Next Article in Journal
Cost-Effective TinyML-Ready Design and Field Deployment of a Solar-Powered Environmental Monitoring Data Collector Using LTE-M Communication
Previous Article in Journal
Applications in Neural and Symbolic Artificial Intelligence
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 16
Issue 7
10.3390/app16073236
Due to scheduled maintenance work on our servers, there may be short service disruptions on this website between 11:00 and 12:00 CEST on March 28th.
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Bioenergy Production and Consumption Prediction: The Best Predictors for the Best Machine Learning Models from Hundreds of Variables

by
Vítor João Pereira Domingues Martinho
1,2
1
School of Agriculture (ESAV), Polytechnic Institute of Viseu (IPV), 3500-631 Viseu, Portugal
2
Centre for Environmental and Marine Studies (CESAM), University of Aveiro, 3810-193 Aveiro, Portugal
Appl. Sci. 2026, 16(7), 3236; https://doi.org/10.3390/app16073236
Submission received: 11 March 2026 / Revised: 21 March 2026 / Accepted: 26 March 2026 / Published: 27 March 2026
(This article belongs to the Special Issue Statistics in Data Science: Latest Methods and Applications)
Browse Figures
Versions Notes

Abstract

The selection of variables that can be used to predict another variable is usually a challenge, considering that national and international databases contain a considerable amount of information and that the literature, in some circumstances, is unclear about the most adjusted predictors and the most accurate models. Without appropriate approaches to select the variables, there are actual risks of considering irrelevant information and ignoring important data in the prediction analysis. Artificial intelligence and, in particular, machine learning methodologies provide interesting support for identifying the most important predictors and the most accurate algorithms. In this way, this research intends to identify the most important variables to predict bioenergy production and consumption and select the most accurate models. For this, statistical information from the FAOSTAT database for the year 2023 was considered. This information was analysed considering machine learning approaches following IBM SPSS Modeler (Version 18.4) procedures. The results obtained indicate that 50% of bioenergy is produced and consumed worldwide by five countries (India, China, the United States of America, Brazil and Ethiopia) and most of this energy comes from firewood (60%). Out of a total of 456 inputs (consideration of this set of FAOSTAT variables is a novelty in the literature), bioenergy production and consumption are mainly explained by fuelwood production, with elasticities of 0.75% and 0.7%, respectively. The explanatory variable “fuelwood production” was identified from the most significant variables found by the machine learning approaches and was subsequently used as an independent variable in linear regressions. XGBoost Linear, XGBoost Tree, Linear, CHAID, Tree-AS, and C&R Tree are the most accurate models (lower relative error) for predicting bioenergy production and consumption worldwide.

1. Introduction

Bioenergy is the energy produced by the conversion of biomass, obtained from organic material such as plants, trees, agroforestry residues and organic waste, through processes of combustion, digestion, fermentation, or gasification, that can be used for biofuels, heat or electricity [1]. Food waste is an example of organic surplus [2] that deserves, in specific cases, more attention from the perspective of energy security than from food security [3]. Sorghum is an example of a cereal often used as a source of biomass to produce bioenergy, which has motivated the scientific community to better understand dimensions related to, for example, its microbial processing [4]. Coconut residue is another source of biomass [5], as well as the sugar industry that is used to produce ethanol and biohydrogen [6]. Nonetheless, there has been no consensus regarding the environmental impacts of sugarcane production among the scientific community [7]. Certain plants that can be considered as a source of biomass may also be used for other purposes, including medicinal and environmental conservation applications [8], what brings challenges in terms of competition for land, for example.
Latin America is a region with particularly good specific conditions to produce biomass and for the generation of bioenergy [9]. Bioenergy appears in the literature that is associated with research fields such as forest bioeconomy, sustainability, global warming mitigation [10], wastewater reuse, soil and water remediation [11] and hydrothermal pretreatment [12]. Some sources of biomass are considered by the literature as unprofitable to produce bioenergy, such as hemp [13]. Other studies found sorghum as less competitive in bioenergy markets [14]. Microalgae can also contribute to the carbon cycle [15] and to obtaining bioenergy [16]. The concept of bioenergy is identified by some researchers as an old theme in some topics (crop planning, for example) [17]. In other topics, bioenergy is an emerging theme [18], including those associated with food waste, particularly in the interrelationships between sustainability and food chains [19].
The literature reveals that there is still a field to be explored on the use of machine learning approaches in bioenergy studies [20], with the artificial neural network being one of the most frequent algorithms [21]. The machine learning algorithms can bring relevant added value, for example, to reduce the risks associated with the uncertainty in the biomass chains [22] and to predict pyrolysis outputs [23].
Considering these gaps in the literature, this research aims to identify the most important predictors and the most accurate algorithms, considering statistical information from the FAOSTAT [24] database and following the procedures proposed by the software IBM SPSS Modeler [25] and Stata (Version 19.0) [26,27]. Regarding machine learning approaches using IBM SPSS Modeler software, in order to reduce the risk of overfitting, considering the size of the dataset, preference was given to the size of the test partition (50%), five-fold cross-validation was employed, and a model was built for each partition. The most relevant variables and the most accurate models were selected by the software, using correlation and relative error scores. The inclusion of 456 FAOSTAT variables as inputs is a novelty, not only because it yields better results than models with fewer variables (that is not the focus of this study) but because it helps to identify which variables are most important for predicting the global bioenergy framework from a vast array of indicators. Furthermore, this research is groundbreaking, as the studies currently available demonstrate the need for an automatic variable selection method, since this helps to resolve the problem of the “curse of dimensionality”, which affects large-scale databases. The machine learning approach presented here complements established econometric methods, demonstrating how the non-linear patterns in energy data can be analysed effectively. Furthermore, existing bioenergy markets face unpredictable price fluctuations, creating an immediate need for advanced forecasting systems that provide accurate results. In short, this study aims to address the following gaps identified in the literature: to identify the most important predictors of bioenergy production and consumption; and to select the most accurate predictive models.

2. Literature Review

The recycling of biomass waste into bioenergy can be a sustainable practice, mitigating, in some circumstances, environmental and health risks. It is estimated that 30–40% of food is lost in the agrifood chains [28], which create challenges and opportunities for reusing this organic material. Nonetheless, more insights are still needed on topics such as soil health implications, crop yields, and the economic viability of composting techniques [29].
The relationships between bioenergy production and food security have motivated the scientific community [30] and should be properly addressed by decision-makers and policymakers [31]. Bioenergy production, in certain cases, may compromise food security, due the competition for production factors, including land and water. Bioenergy fields also deserve the attention of researchers, particularly in issues related to the wood chain [32].
These new technologies offer a set of innovative solutions to deal with the challenges associated with food and energy security [33]. It is important, however, to better understand the impacts of using bioenergy crops on agricultural systems and agrifood chains [34]. Other promising technologies related to bioenergy framework may have relevant added value for sustainable development, such as microbial fuel cells [35] and microbial enzymes [36]. There are several sources of biomass that still need to be better addressed, specifically to highlight their environmental and economic benefits [37].
Machine learning and emerging approaches may contribute significantly to the domains associated with the bioenergy production and consumption, nonetheless more studies are needed to identify the most accurate methods [38]. Artificial intelligence, namely machine learning, may provide important contributions for livestock waste management [39], the integration of renewable energy to power systems [40], harmonising approaches in agroforestry waste valorisation [41], the optimisation of processes [42], waste collection and transport [43], and dealing with current sustainability challenges [44] and energy efficiency [45]. These advances create opportunities to deal with the increased demand for energy, mitigate environmental impacts, and reduce biodiversity losses [46]. Biotechnology has additional potential to promote more sustainability in bioenergy/biorefinery contexts [47], principally improving plant resilience to biotic and abiotic stresses [48].
The consideration of some methodologies for bioenergy assessment still needs to be standardised, namely in terms of system boundaries, to improve the conditions of benchmarking between different studies [49]. The life-cycle sustainability assessment approaches for assessing the alternative use and processing of biomass still lack consideration [50].
Biomass production and use have positive and negative impacts on the environment [51] and climate change [52], depending on the production options [53], management decisions [54] and conversion practices [55], and this calls for adjusted land management strategies, policies and life-cycle assessments [56]. Bioenergy is a clean resource [57] and may contribute to reducing greenhouse gas emissions [58]. But for an effective contribution of bioenergy to sustainability, several factors must be considered, such as the location of conversion plants, namely because the transport logistics have impacts on the economic and environmental dimensions [59].
The European Union renewable energy legislation indicates that the impacts on environmental and ecosystem services from bioenergy production should be assessed [60], suggesting that it is not guaranteed that the bioenergy frameworks are always free of negative consequences on sustainability. Climate-smart agriculture practices may have a relevant contribution to a more sustainable development [61] due to their added value to improving the efficiency of farming activities. Despite the concerns regarding impacts of bioenergy production and use on sustainability, only a residual part of the studies realised consider the three dimensions [62].
The valorisation of agroforestry biomass has several drivers and barriers. The main barriers are associated with unadjusted public and private strategies, insufficient stakeholder skills, investment constraints and production costs. The drivers are related to the new technologies of conversion and public policies [63].

3. Data Analysis

The biggest bioenergy producers are India, China, the United States of America and Brazil, which together represent almost 50% of the world’s production (Figure 1 and Table 1). India has the highest production (8,605,125 Tj) which represents 17% of the world’s bioenergy produced. Figure 1 shows some production clusters around China and India, Central Africa, Germany, Brazil and the United States. On the other hand, fuelwood represents 60% of the total bioenergy produced worldwide, followed by other vegetal material and residues. Charcoal, for example, another forestry related product represents 4% of all bioenergy production (Table 2).
In general, bioenergy consumption follows a global geographical distribution similar to that observed for bioenergy production. India, China, the United States of America and Brazil consume almost 50% of the world’s bioenergy. Of the 39,473,209 Tj of bioenergy consumed, 7,612,473 Tj (19%) was consumed in India and 4,664,906 Tj (12%) in China (Figure 2 and Table 3). Fuelwood and other vegetal material and residues represent almost two-thirds of the energy sources of global bioenergy consumption (Table 4).

4. Results

This section will have two subsections for bioenergy production and bioenergy consumption, respectively. In each case, several variables were considered as inputs for the machine learning models. The statistical information for these variables was obtained from the FAOSTAT database. Considering the organisation of this database, the following groups of information, related to social, economic and environmental dimensions, were considered (with a total of 457 variables): emissions from crops; emissions from forests; emissions from pre and post agricultural production; emissions totals; employment indicators from agriculture and agrifood systems; cropland nutrient balance; temperature change on land; food security indicators; food balances; forestry production and trade; fertilisers by nutrient; land use elements; investment government expenditure; macro indicators; annual population; producer prices; crops and livestock production; SDG indicators; food and diet supply utilisation accounts; crop and livestock trade; trade indices; and value shares by industry and primary factors.

4.1. Bioenergy Production

Table 5 reveals that XGBoost Linear, CHAID, Tree-AS, Linear and XGBoost Tree are the most accurate (considering the relative error) algorithms to predict total bioenergy production. XGBoost Linear is the most accurate model for the testing and training (Table 6) set. The correlation is also high for this algorithm. It should be noted that the correlations obtained for some models in the training set are relatively high, particularly for XGBoost Linear, but then drop slightly in the testing set, indicating that the model has good generalisability. The Tree-AS model tends to have lower accuracy scores because it avoids overfitting, uses simpler algorithms and prioritises generalisation.
The relationships between the observed values and predicted ones presented in Figure 3 confirm the accuracy of the XGBoost Linear, CHAID, Tree-AS, Linear and XGBoost Tree models. This figure suggests the existence of some outliers at the top. The most important predictors (identified taking into account all the models shown in the respective tables) are related to different types of bioenergy consumption (biogasoline, solid biofuels, fuelwood and biodiesel), including total bioenergy consumption, total emissions of CO2 from forestland, N2O emissions from burning crop residues of maize, CH4 emissions from burning crop residues of rice, N2O emissions from pesticides manufacturing, and CH4 emissions from manure management (Figure 4). Variables such as emissions from forestland or the burning of crop residues are important predictors of biomass production because they reflect biomass availability, are associated with agricultural and forestry activities, indicate energy potential, and reveal structural patterns in economic dynamics.
To minimise potential redundancy issues, the results presented in the following tables and figures of this section were obtained without considering input variables related to bioenergy production and bioenergy consumption. With these new inputs, the most accurate models are presented in Table 7 and Table 8 for the testing and training set. In these tables, SGBoost Tree appears among the most accurate algorithms. This model has the lowest relative error and the highest correlation for the testing set. It also presents good findings for the correlation and relative error in the training set.
The good accuracy (with some outliers) is also confirmed by the relationships among the observed values and the predicted results presented in Figure 5. The presence of outliers in Figure 3 and Figure 5 is likely due to the fact that four to five of the countries account for approximately 50% of global bioenergy production. A sensitivity analysis shows that excluding these countries alters the results, but the aim of this study is to provide insights into the global situation as a whole. It would be worthwhile in future research to conduct predictive analyses by country group. The most important predictors in this approach are the following (Figure 6): crop emissions from burning crop residues; production of wood fuel; domestic supply quantity of cottonseed oil; emissions of N2O from burning crop residues of maize; emissions of CO2 from all sectors with LULUCF; domestic supply quantity of rice and products; emissions of N2O from livestock; domestic supply of alcoholic beverages; emissions of N2O from energy use; and emissions of N2O from all sectors with LULUCF.
Considering the results obtained for the most important predictors of bioenergy production (with and without inputs related to bioenergy production and consumption) the following variables were considered to carry out a linear regression to quantify the relationships between the most important predictive variables and the total energy output: crop emissions from burning crop residues; production of wood fuel; and domestic supply quantity of cottonseed oil. To obtain a linear model, the variables were linearised through logarithms, and the summary statistics are those presented in Table 9. Following the results presented before, Table 10 shows the results obtained for a linear regression with robust standard error (to deal with heteroskedasticity). The use of robust standard errors helps to address the problem of heteroscedasticity, making significance tests more reliable, even when the variance of the residuals is not constant across observations. This model was built considering the findings for the most important predictors and to deal with the problems of multicollinearity (VIF test), but the RESET test confirms the adequacy of the model. These results highlight that, on average, countries with an additional 1% the production of wood fuel tend to have 0.75% of additional bioenergy production.

4.2. Bioenergy Consumption

XGBoost Tree and XGBoost Linear are the most accurate models for testing set (Table 11), both in terms of relative error (0.230 and 0.378, respectively) and correlation (0.942 and 0.852, respectively). For the training set (Table 12), XGBoost Linear appears with the best accuracy (0.002 for relative error and 0.999 for correlation).
Figure 7 also shows that there is a good accuracy of the models considered, considering the relationships between the observed and predicted values for bioenergy consumption, with some outliers. In general, the results obtained are similar to those found for bioenergy production. The most important predictors are related to the different types of bioenergy production, emissions of CH4 from agrifood systems, N2O emissions from burning crop residues of maize, CH4 emissions from enteric fermentation and the domestic supply of rice and products (Figure 8).
Taking into account bioenergy consumption as a target and the remaining variables as inputs, with the exception of the different types of bioenergy production and consumption, the most accurate models are those presented in Table 13 and Table 14, respectively, for the testing and training set. The linear model, for example, appears as the most accurate for the training set with a relative error of 0.005 and a correlation of 0.997.
Figure 9 reveals interesting accuracy, with some outliers, and Figure 10 presents the most important predictors that in this case are the following: crop emissions from burning crop residues; emissions of CH4 from agrifood systems; production of wood fuel; emissions of CH4 from enteric fermentation; emissions of CH4 from manure management; burning crop residues emissions of N2O from maize; rural population; area harvest of rice; burning crop residues from maize; domestic supply quantity of alcoholic beverages.
Considering the results obtained for the most important predictors of bioenergy consumption, the variables presented in Table 15 and Table 16 were selected for a linear regression. The summary statistics are presented in Table 15 and the results for a linear regression with robust standard error are those highlighted in Table 16. The regression results reveal that when crop emissions from burning crop residues, emissions of CH4 from agrifood systems, and production of wood fuel increase by 1%, the total bioenergy consumption increases by 0.1%, 0.2% and 0.7%, respectively.

5. Discussion

There is a wide variety of organic matter (biomass) that can be used to produce bioenergy. This renewable and sustainable energy is generated by adjusted processes and is used for different purposes [1,2,4,5,6]. There is no consensus among the scientific community on whether some sources of bioenergy are sustainable [7], including in terms of competition for resources [8]. Another concern relates to the profitability of bioenergy production [13] and the competitiveness of the respective markets [14]. Further research is needed to better understand the environmental impacts, agricultural implications and economic viability of some processes [29]. Latin America is an important player in the international bioenergy framework [9]. New technologies bring significant added value, particularly in addressing the links between food and energy security [33]. Specifically on machine learning methodologies, further insights are needed to select the most accurate models [38]. Animal waste management [39], renewable energy integration [40] and energy efficiency [45] are examples where artificial intelligence can make interesting contributions. On the other hand, the contributions of life-cycle sustainability assessments for analysing biomass conversion could be further explored [50]. Public policies have important implications in these contexts [63].
The production and consumption of bioenergy follow a similar global geographical distribution, with India, China, the United States of America and Brazil being the biggest players on the international stage, accounting for almost 50% of the sector. On the other hand, fuelwood accounts for almost 60% of the world’s bioenergy produced and consumed, demonstrating its importance in these contexts.
Considering a total of 456 inputs for machine learning approaches, the most important predictors of bioenergy production are variables related to bioenergy consumption, CO2 emissions, N20 emissions, CH4 emissions, wood fuel production, and the domestic supply of cottonseed oil, rice and alcoholic beverages. Emissions come from a wide variety of sources, including forestland, burning crop residues of maize and rice, pesticide manufacturing, and manure management. The most accurate algorithms are XGBoost Linear, CHAID, Tree-AS, Linear, and XGBoost Tree. A linear regression analysis shows that bioenergy production is mainly explained by wood fuel production, with an elasticity of 0.75% (for a 1% variation in the independent variable). In assessing bioenergy consumption, the most accurate algorithms are almost the same, with CHAID replaced by C&R Tree. The most important predictors are related to bioenergy production, CH4 emissions (agrifood systems), N2O emissions (burning crop residues of maize), CH4 emissions (enteric fermentation and manure management), the domestic supply of rice, production of wood fuel, rural population, and area harvested for rice. For bioenergy consumption, CH4 emissions from agrifood systems and wood fuel production are the explanatory variables with statistical significance in linear regression, with elasticities of 0.2% and 0.7% (for a 1% variation) respectively.
The results for the predictors show that the approaches considered captured the dynamics of associated processes, such as the burning of agricultural and forestry residues, agricultural and livestock management, fuelwood production, and biomass management. In this way, the machine learning algorithms captured complex and non-linear relationships between social, economic, environmental, and energy variables. Global strategies for carbon neutrality will have an impact on the identified predictors and, consequently, on the bioenergy market.

6. Conclusions

In terms of practical implications, it should be noted that 50% of the world’s bioenergy is produced and consumed by five countries (India, China, the United States, Brazil and Ethiopia) and much of this bioenergy comes from fuelwood (60%). The production and consumption of bioenergy have a very similar geographical distribution worldwide. The most important predictors (found through machine learning approaches) of bioenergy production and consumption worldwide (out of a total of 456 inputs) are related to greenhouse gas emissions (from agrifood systems, the burning of maize and rice residues, and manure management), fuelwood production, domestic supply of rice and alcoholic beverages, rural population, and harvested rice area. On the other hand, fuelwood production accounts for 70% and 75%, respectively, of marginal bioenergy consumption and production. These results show the interconnection between bioenergy production and consumption; these variables highlight the importance of some sectors for bioenergy markets, but also the possible environmental impacts of some practices of production and consumption of this energy source. In addition to the metrics, the results obtained provide valuable insights for decision-making processes, stakeholders and policymakers. Priority should be given to more sustainable processes for the production and consumption of bioenergy, improving waste management, promoting greater sustainability in biomass supply chains, and mitigating emissions from bioenergy technologies. For policy recommendations, it is suggested to create internationally standardised methodologies (particularly those related to life-cycle sustainability assessment) that enable real impacts on the sustainability of bioenergy production and use to be assessed. For future research, it would be important to further explore the relationships between the predictors found and bioenergy production and consumption. Priority should also be given to dynamic feature selection, as it improves forecasting accuracy and maintains the model’s interpretability in the face of changing policies, markets and global economic conditions.

Funding

This work was funded by national funds through FCT—Fundação para a Ciência e a Tecnologia I.P., under the project CESAM-Centro de Estudos do Ambiente e do Mar, references UID/50017/2025 (https://doi.org/10.54499/UID/50017/2025) and LA/P/0094/2020 (https://doi.org/10.54499/LA/P/0094/2020).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.

Acknowledgments

Furthermore we would like to thank the Polytechnic Institute of Viseu for their support.

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Abioye, K.J.; Harun, N.Y.; Chofreh, A.G.; Kamyab, H.; Yusuf, M.; Jagaba, A.H.; Balogun, A.O.; Ayodele, B.V.; Ighalo, J.O. Trends and Advancements in Utilization of Biomass Waste for Gasification: A Bibliometric Review. GCB Bioenergy 2025, 17, e70095. [Google Scholar] [CrossRef]
  2. Elkhwesky, Z.; Gumbo, A.R.; Abuelhassan, A.E.; Hassan, H.; Elkhwesky, F.F.Y.; El Manzani, Y. A Systematic Review with Bibliometric Analysis of Food Waste Research in Contemporary Hospitality. J. Culin. Sci. Technol. 2025, 1–33. [Google Scholar] [CrossRef]
  3. Łaba, S.; Olech, I. Food waste in the context of Poland’s food security: Identifying research path dependency through bibliometric research. Bulg. J. Agric. Sci. 2023, 29, 140–146. [Google Scholar]
  4. Adnan, A.; Hastuti, C.O.I.; Hadiarto, A.; Lestari, I.P.; Haryati, Y.; Indrasti, R.; Cahyaningrum, H.; Qomariyah, N.; Permana, D.; Nurmalinda; et al. Evolving Paradigms in Sorghum Research: A Bibliometric and Content Analysis of Global Trends and Future Directions. Int. J. Des. Nat. Ecodyn. 2024, 19, 887–898. [Google Scholar] [CrossRef]
  5. Vieira, F.; Santana, H.E.P.; Silva, D.P.; Ruzene, D.S. A Bibliometric Description of Organosolv Pretreatment for Coconut Waste Valorization. Bioenergy Res. 2023, 16, 2115–2130. [Google Scholar] [CrossRef]
  6. Aguilar-Rivera, N.; Enríquez-Poy, M. Strategies for the Development of Sugarcane Ecotourism in Veracruz, Mexico. Sugar Tech 2025. [Google Scholar] [CrossRef]
  7. Dos Santos, E.R.; Carvalho, W.D.; Mustin, K. Bibliometric Analysis of Global Research on Sugarcane Production and Its Effects on Biodiversity: Trends, Critical Points, and Knowledge Gaps. Conservation 2025, 5, 67. [Google Scholar] [CrossRef]
  8. Aminah, A.; Adnan, A.; Qomariyah, N.; Krisnan, R.; Anggraini, L.; Supatmi, S.; Danu, D.; Suita, E.; Cahyono, D.D.N.; Lukman, A.H.; et al. Unlocking the Potential of Pongamia pinnata: A Scoping Review and Bibliometric Analysis (ScoRBA) Guided by the PAGER Framework. For. Sci. Technol. 2025, 21, 302–317. [Google Scholar] [CrossRef]
  9. Ascencio-Galván, M.L.; López-Agudelo, V.A.; Gómez-Ríos, D.; Ramirez-Malule, H. A bibliometric landscape of polyhydroxyalkanoates production from low-cost substrates by Cupriavidus necator and its perspectives for the Latin American bioeconomy. J. Appl. Biol. Biotechnol. 2024, 12, 48–61. [Google Scholar] [CrossRef]
  10. Biancolillo, I.; Paletto, A.; Bersier, J.; Keller, M.; Romagnoli, M. A literature review on forest bioeconomy with a bibliometric network analysis. J. For. Sci. 2020, 66, 265–279. [Google Scholar] [CrossRef]
  11. Jia, Y.-X.; Xu, J.-Z.; Li, Y.-W.; Liu, X.-Y.; Xu, Y.; Wei, Q.; Hu, Z.-W.; Tian, W. Analysis of research hotspots and development trends of iron-modified biochar based on bibliometrics. J. Ecol. Rural Environ. 2025, 41, 569–578. [Google Scholar] [CrossRef]
  12. Yu, Y.; Pei, W.; Zhao, X.; Cárdenas-Oscanoa, A.J.; Huang, C. Global evolution of research on autohydrolysis (hydrothermal) pretreatment as a green technology for biorefineries: A bibliometric analysis. J. Bioresour. Bioprod. 2025, 10, 92–110. [Google Scholar] [CrossRef]
  13. Budhathoki, R.; Maraseni, T.; Apan, A. Enviro-economic and feasibility analysis of industrial hemp value chain: A systematic literature review. GCB Bioenergy 2024, 16, e13141. [Google Scholar] [CrossRef]
  14. George, T.T.; Obilana, A.O.; Oyenihi, A.B.; Obilana, A.B.; Akamo, D.O.; Awika, J.M. Trends and progress in sorghum research over two decades, and implications to global food security. S. Afr. J. Bot. 2022, 151, 960–969. [Google Scholar] [CrossRef]
  15. Zhong, C.; Li, T.; Bi, R.; Sanganyado, E.; Huang, J.; Jiang, S.; Zhang, Z.; Du, H. A systematic overview, trends and global perspectives on blue carbon: A bibliometric study (2003–2021). Ecol. Indic. 2023, 148, 110063. [Google Scholar] [CrossRef]
  16. Li, M.; Wang, Y.; Zhang, J.; Liu, B.; Xue, H.; Wu, L.; Li, Z. Knowledge Mapping of High-Rate Algal Ponds Research. Water 2023, 15, 1916. [Google Scholar] [CrossRef]
  17. Deo, A.; Sawant, N.; Arora, A.; Karmakar, S. How has scientific literature addressed crop planning at farm level: A bibliometric-qualitative review. Farming Syst. 2025, 3, 100139. [Google Scholar] [CrossRef]
  18. Umunnawuike, C.; Mahat, S.Q.A.; Nwaichi, P.I.; Money, B.; Agi, A. Biohydrogen production for sustainable energy transition: A bibliometric and systematic review of the reaction mechanisms, challenges, knowledge gaps and emerging trends. Biomass Bioenergy 2024, 188, 107345. [Google Scholar] [CrossRef]
  19. Prasetiyo, B.D.; Sholihah, Q.; Riniwati, H.; Wardana, F.Y. Unraveling Factors Affecting Food Waste Globally: A PRISMA-Based Systematic Literature Review and Bibliometric Analysis. Int. J. Des. Nat. Ecodyn. 2024, 19, 2213–2230. [Google Scholar] [CrossRef]
  20. Helal, M.A.; Anderson, N.; Wei, Y.; Thompson, M. A Review of Biomass-to-Bioenergy Supply Chain Research Using Bibliometric Analysis and Visualization. Energies 2023, 16, 1187. [Google Scholar] [CrossRef]
  21. Cheng, Y.; Zhao, C.; Neupane, P.; Benjamin, B.; Wang, J.; Zhang, T. Applicability and Trend of the Artificial Intelligence (AI) on Bioenergy Research between 1991–2021: A Bibliometric Analysis. Energies 2023, 16, 1235. [Google Scholar] [CrossRef]
  22. Kumar, G.; Sowlati, T. Hybrid Simulation Optimization Method for Biomass Supply Chain Planning: A Systematic Review. Bioenergy Res. 2025, 18, 101. [Google Scholar] [CrossRef]
  23. Laouidi, K.; Habchi, S.; Fanezoune, C.K.; Sallek, B.; Kussul, N.; El Bari, H. Advancements in artificial neural networks and fast pyrolysis of biomass processing: A comprehensive review and a bibliometric analysis. J. Anal. Appl. Pyrolysis 2025, 189, 107098. [Google Scholar] [CrossRef]
  24. FAOSTAT Several Statistics. 2026. Available online: https://www.fao.org/faostat/en/#data (accessed on 11 January 2026).
  25. IBM SPSS, Version 18.4; IBM SPSS Modeler. IBM: Armonk, NY, USA, 2026.
  26. StataCorp. Stata 19 Base Reference Manual 2025; StataCorp: College Station, TX, USA, 2025. [Google Scholar]
  27. StataCorp. Stata Statistical Software, Release 19; StataCorp: College Station, TX, USA, 2025. [Google Scholar]
  28. Peguero, D.A.; Gold, M.; Vandeweyer, D.; Zurbrügg, C.; Mathys, A. A Review of Pretreatment Methods to Improve Agri-Food Waste Bioconversion by Black Soldier Fly Larvae. Front. Sustain. Food Syst. 2022, 5, 745894. [Google Scholar] [CrossRef]
  29. Abdellah, Y.A.Y.; Yang, S.; Yang, C.; Yang, Z.; Elsheikh, E.A.; Keiblinger, K.M.; Mohamed, T.A.; Rana, M.S.; Liu, D.; Yu, F. Tobacco waste valorization through composting: A systematic review of biomass conversion efficiency and circular bioeconomy strategies. Biomass Bioenergy 2025, 201, 108137. [Google Scholar] [CrossRef]
  30. Muscat, A.; De Olde, E.M.; de Boer, I.J.M.; Ripoll-Bosch, R. The battle for biomass: A systematic review of food-feed-fuel competition. Glob. Food Secur. 2020, 25, 100330. [Google Scholar] [CrossRef]
  31. Ahmed, S.; Warne, T.; Smith, E.; Goemann, H.; Linse, G.; Greenwood, M.; Kedziora, J.; Sapp, M.; Kraner, D.; Roemer, K.; et al. Systematic review on effects of bioenergy from edible versus inedible feedstocks on food security. Npj Sci. Food 2021, 5, 9. [Google Scholar] [CrossRef]
  32. Auer, V.; Rauch, P. Wood supply chain risks and risk mitigation strategies: A systematic review focusing on the Northern hemisphere. Biomass Bioenergy 2021, 148, 106001. [Google Scholar] [CrossRef]
  33. Barbosa Júnior, M.R.; Moreira, B.R.A.; de Brito Filho, A.L.; Tedesco, D.; Shiratsuchi, L.S.; Silva, R.P. UAVs to Monitor and Manage Sugarcane: Integrative Review. Agronomy 2022, 12, 661. [Google Scholar] [CrossRef]
  34. Cumplido-Marin, L.; Graves, A.R.; Burgess, P.J.; Morhart, C.; Paris, P.; Jablonowski, N.D.; Facciotto, G.; Bury, M.; Martens, R.; Nahm, M. Two novel energy crops: Sida hermaphrodita (L.) rusby and Silphium perfoliatum L.-state of knowledge. Agronomy 2020, 10, 928. [Google Scholar] [CrossRef]
  35. Dixit, S.; Imam, A.; Rajput, R.S.; Rajput, M.S. Microbial fuel cells as future of green technology: A systematic review. 3 Biotech 2025, 15, 218. [Google Scholar] [CrossRef] [PubMed]
  36. Golgeri, M.D.B.; Mulla, S.I.; Bagewadi, Z.K.; Tyagi, S.; Hu, A.; Sharma, S.; Bilal, M.; Bharagava, R.N.; Ferreira, L.F.R.; Gurumurthy, D.M.; et al. A systematic review on potential microbial carbohydrases: Current and future perspectives. Crit. Rev. Food Sci. Nutr. 2024, 64, 438–455. [Google Scholar] [CrossRef] [PubMed]
  37. Ekawati, I.; Syabana, R.A.; Isdiantoni, I.; Kountjoro, M.P.; Prasetyo, E.N.; Patmawati, P.; Khan, A.S.; Hasfiah, H. Exploring the Diverse Benefits of Pine Needle Litter for Environmental Sustainability and Economic Development: A Systematic Literature Review. Sarhad J. Agric. 2023, 39, 153–162. [Google Scholar] [CrossRef]
  38. Fassnacht, F.; Hartig, F.; Latifi, H.; Berger, C.; Hernández, J.; Corvalán, P.; Koch, B. Importance of sample size, data type and prediction method for remote sensing-based estimations of aboveground forest biomass. Remote Sens. Environ. 2014, 154, 102–114. [Google Scholar] [CrossRef]
  39. Grieco, R.; Cervelli, E.; Bovo, M.; Pindozzi, S.; di Perta, E.S.; Tassinari, P.; Torreggiani, D. The role of geospatial technologies for sustainable livestock manure management: A systematic review. Sci. Total Environ. 2024, 954, 176687. [Google Scholar] [CrossRef]
  40. Masood, A.; Ahmed, U.; Hassan, S.Z.; Khan, A.R.; Mahmood, A. Economic Value Creation of Artificial Intelligence in Supporting Variable Renewable Energy Resource Integration to Power Systems: A Systematic Review. Sustainability 2025, 17, 2599. [Google Scholar] [CrossRef]
  41. Mehdizadeh, M.; Omidi, A.; Matindike, R.; Nigussie, Z.G.; Ikegwu, T.; Agu, H.O.; Onen, P.; Gumula, I.; Ezzat, S.M.; Merghany, R.M. Agri-Waste Valorization: Pathways to Sustainable Bioenergy and Biochemical Innovation. Circ. Econ. Sustain. 2025, 5, 5247–5277. [Google Scholar] [CrossRef]
  42. Shahverdi, N.; Saffari, A.; Amiri, B. A systematic review of artificial intelligence and machine learning in energy sustainability: Research topics and trends. Energy Rep. 2025, 13, 5551–5578. [Google Scholar] [CrossRef]
  43. Workie, E.; Kumar, V.; Bhatnagar, A.; He, Y.; Dai, Y.; Tong, Y.W.; Peng, Y.; Zhang, J.; Fu, C. Advancing the bioconversion process of food waste into methane: A systematic review. Waste Manag. 2023, 156, 187–197. [Google Scholar] [CrossRef]
  44. Khan, S.; Mim, J.J.; Shorna, J.F.; Hasan, A.; Tarek, H.R.; Islam, A.; Hossain, N. Machine learning for renewable energy advancements: Prospects and emerging techniques. Energy Rep. 2026, 15, 109008. [Google Scholar] [CrossRef]
  45. Postawa, K.; Buczyński, K.; Maj, G.; Klimek, K.E.; Kapłan, M. Enhanced artificial neural networks in predicting energy efficiency and minimize emissions from thermal treatment of lignocellulosic biowaste. Biomass Bioenergy 2026, 206, 108638. [Google Scholar] [CrossRef]
  46. Grace, O.M.; Lovett, J.C.; Gore, C.J.N.; Moat, J.; Ondo, I.; Pironon, S.; Langat, M.K.; Pérez-Escobar, O.A.; Ross, A.; Abbo, M.S.; et al. Plant Power: Opportunities and challenges for meeting sustainable energy needs from the plant and fungal kingdoms. Plants People Planet 2020, 2, 446–462. [Google Scholar] [CrossRef]
  47. Meneses-Montero, K.; Rojas-Villalta, D.; Orozco-Ortiz, C.; Jerez-Navarro, A.; Goméz-Espinoza, O. Efforts on Small- and Large-Scale Cultivation of Isochrysis galbana for Enhanced Growth and Lipid Production: A Systematic Review Towards Biorefinery Applications. Bioenergy Res. 2025, 18, 37. [Google Scholar] [CrossRef]
  48. Longhi Cervantes, D.S.; Leal, G.M.; da Silva Fortirer, J.; de Oliveira, L.F.; Navarro, B.V.; Buckeridge, M.S. microRNAs and stress adaptation in grasses: A systematic review. Plant Physiol. Biochem. 2026, 230, 110783. [Google Scholar] [CrossRef]
  49. Musule, R.; Bonales-Revuelta, J.; Mwampamba, T.H.; Gallardo-Álvarez, R.M.; Masera, O.; García-Bustamante, C.A. Life Cycle Assessment of Forest-Derived Solid Biofuels: A Systematic Review of the Literature. Bioenergy Res. 2022, 15, 1711–1732. [Google Scholar] [CrossRef]
  50. Ramos-Reyes, M.F.; González-López, M.E.; Barajas-Álvarez, P.; Garcia-Garcia, C.E.; Tuesta-Popolizio, D.A.; Mussatto, S.I.; Gradilla-Hernández, M.S. Exploring the potential of distillery vinasses through upcycling: Pathways to a circular economy. Environ. Technol. Innov. 2025, 38, 104072. [Google Scholar] [CrossRef]
  51. Harris, Z.M.; Spake, R.; Taylor, G. Land use change to bioenergy: A meta-analysis of soil carbon and GHG emissions. Biomass Bioenergy 2015, 82, 27–39. [Google Scholar] [CrossRef]
  52. Robledo-Abad, C.; Althaus, H.; Berndes, G.; Bolwig, S.; Corbera, E.; Creutzig, F.; Garcia-Ulloa, J.; Geddes, A.; Gregg, J.S.; Haberl, H.; et al. Bioenergy production and sustainable development: Science base for policymaking remains limited. GCB Bioenergy 2017, 9, 541–556. [Google Scholar] [CrossRef]
  53. Sharma, S.; Chaubey, I. Surface and subsurface transport of nitrate loss from the selected bioenergy crop fields: Systematic review, analysis and future directions. Agriculture 2017, 7, 27. [Google Scholar] [CrossRef]
  54. Udali, A.; Chung, W.; Talbot, B.; Grigolato, S. Managing harvesting residues: A systematic review of management treatments around the world. Forestry 2025, 98, 117–135. [Google Scholar] [CrossRef]
  55. Knight, D.R.; Goldsworthy, M.; Smith, P. Are biomass feedstocks sustainable? A systematic review of three key sustainability metrics. GCB Bioenergy 2024, 16, e13187. [Google Scholar] [CrossRef]
  56. Freer-Smith, P.; Bailey-Bale, J.H.; Donnison, C.L.; Taylor, G. The good, the bad, and the future: Systematic review identifies best use of biomass to meet air quality and climate policies in California. GCB Bioenergy 2023, 15, 1312–1328. [Google Scholar] [CrossRef]
  57. Lungaho, M.; Ojuederie, O.B.; Odozi, E.B.; Mshelmbula, B.P.; Onawo, L.O.; Igiebor, F.A.; Uselu, A.; Adegboyega, T.T.; Ikhajiagbe, B. From discard to resource: Unlocking the environmental and nutritional value of Bambara groundnut waste. Front. Sustain. Food Syst. 2025, 9, 1684699. [Google Scholar] [CrossRef]
  58. Jebari, A.; Pereyra-Goday, F.; Kumar, A.; Collins, A.L.; Jordana Rivero, M.J.; McAuliffe, G.A. Feasibility of mitigation measures for agricultural greenhouse gas emissions in the UK. A systematic review. Agron. Sustain. Dev. 2024, 44, 2. [Google Scholar] [CrossRef]
  59. Ibarra-Esparza, F.E.; González-López, M.E.; Senés-Guerrero, C.; Chong, J.P.J.; Forrester, S.; Gradilla-Hernández, M.S. Anaerobic co-digestion of agro-industrial wastes: A systematic review focused on feedstock physicochemical parameter optimization. Biomass Bioenergy 2026, 206, 108671. [Google Scholar] [CrossRef]
  60. Meyer, M.A.; Leckert, F.S. A systematic review of the conceptual differences of environmental assessment and ecosystem service studies of biofuel and bioenergy production. Biomass Bioenergy 2018, 114, 8–17. [Google Scholar] [CrossRef]
  61. Ntawuruhunga, D.; Ngowi, E.E.; Mangi, H.O.; Salanga, R.J.; Shikuku, K.M. Climate-smart agroforestry systems and practices: A systematic review of what works, what doesn’t work, and why. For. Policy Econ. 2023, 150, 102937. [Google Scholar] [CrossRef]
  62. Padilla-Rivera, A.; Paredes-Figueroa, M.G.; Güereca, L.P. A systematic review of the sustainability assessment of bioenergy: The case of gaseous biofuels. Biomass Bioenergy 2019, 125, 79–94. [Google Scholar] [CrossRef]
  63. Rijal, P.; Carvalho, H.; Matias, J.; Garrido, S.; Pimentel, C. Drivers and barriers of residual agroforestry biomass valorization: A systematic literature review. Agrofor. Syst. 2025, 99, 81. [Google Scholar] [CrossRef]
Applsci 16 03236 g001
Figure 1. World total bioenergy production (Tj), in 2023.
Figure 1. World total bioenergy production (Tj), in 2023.
Applsci 16 03236 g001
Applsci 16 03236 g002
Figure 2. Total world bioenergy consumption (Tj), in 2023.
Figure 2. Total world bioenergy consumption (Tj), in 2023.
Applsci 16 03236 g002
Applsci 16 03236 g003
Figure 3. Relationships between the observed values and predicted ones, for bioenergy production, in 2023.
Figure 3. Relationships between the observed values and predicted ones, for bioenergy production, in 2023.
Applsci 16 03236 g003
Applsci 16 03236 g004
Figure 4. The most important predictors for bioenergy production in 2023. Note: Bioenergy BIO-CON-Biogasoline (bioenergy consumption-biogasoline, Tj); Bioenergy BIO-CON-TOB (total bioenergy consumption, Tj); Bioenergy BIO-CON-SOB (bioenergy consumption-solid biofuels, Tj); Bioenergy BIO-CON-Fuelw (bioenergy consumption-fuelwood, Tj); Emissions Total EMT-ECO-Fores (emissions CO2-forestland, kt); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Bioenergy BIO-CON-Biod (energy consumption-biodiesel, Tj); Crop Emission CRO-BEC-Rice (burning crop residues-emissions CH4-rice, kt); Emissions Pre and Pro EPP-ENO-PEM (emissions N2O-pesticides manufacturing, kt); Emissions Total EMT-ECH-MAM (emissions CH4-manure management, kt).
Figure 4. The most important predictors for bioenergy production in 2023. Note: Bioenergy BIO-CON-Biogasoline (bioenergy consumption-biogasoline, Tj); Bioenergy BIO-CON-TOB (total bioenergy consumption, Tj); Bioenergy BIO-CON-SOB (bioenergy consumption-solid biofuels, Tj); Bioenergy BIO-CON-Fuelw (bioenergy consumption-fuelwood, Tj); Emissions Total EMT-ECO-Fores (emissions CO2-forestland, kt); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Bioenergy BIO-CON-Biod (energy consumption-biodiesel, Tj); Crop Emission CRO-BEC-Rice (burning crop residues-emissions CH4-rice, kt); Emissions Pre and Pro EPP-ENO-PEM (emissions N2O-pesticides manufacturing, kt); Emissions Total EMT-ECH-MAM (emissions CH4-manure management, kt).
Applsci 16 03236 g004
Applsci 16 03236 g005
Figure 5. Relationships between the observed values and predicted ones, for bioenergy production (without inputs for different types of bioenergy production and bioenergy consumption), in 2023.
Figure 5. Relationships between the observed values and predicted ones, for bioenergy production (without inputs for different types of bioenergy production and bioenergy consumption), in 2023.
Applsci 16 03236 g005
Applsci 16 03236 g006
Figure 6. The most important predictors for bioenergy production (without inputs for different types of bioenergy production and bioenergy consumption), in 2023. Note: Crop Emissions CRO-BBD-ALC (burning crop residues-biomass burned dry matter-all crops, t); Forestry FOR-PRO-WOFU (production-wood fuel, m3); Food Balance FOB-DSQ-COTOI (domestic supply quantity-cottonseed oil, 1000 t); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Emissions Total EMT-ECO-ASL (emissions CO2-all sectors with LULUCF); Food Balance FOB-DSQ-RIaP (domestic supply quantity-rice and products, 1000 t); Emissions Total EMT-ENO-EFL (emissions N2O-emissions from livestock, kt); Food Balance FOB-DSQ-ALBE (domestic supply-alcoholic beverages, 1000 t); Emissions Pre and Pro EPP-ENO-ENU (emissions N2O-energy use, kt); Emissions Total EMT-ENO-ASL (emissions N2O-all sectors with LULUCF).
Figure 6. The most important predictors for bioenergy production (without inputs for different types of bioenergy production and bioenergy consumption), in 2023. Note: Crop Emissions CRO-BBD-ALC (burning crop residues-biomass burned dry matter-all crops, t); Forestry FOR-PRO-WOFU (production-wood fuel, m3); Food Balance FOB-DSQ-COTOI (domestic supply quantity-cottonseed oil, 1000 t); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Emissions Total EMT-ECO-ASL (emissions CO2-all sectors with LULUCF); Food Balance FOB-DSQ-RIaP (domestic supply quantity-rice and products, 1000 t); Emissions Total EMT-ENO-EFL (emissions N2O-emissions from livestock, kt); Food Balance FOB-DSQ-ALBE (domestic supply-alcoholic beverages, 1000 t); Emissions Pre and Pro EPP-ENO-ENU (emissions N2O-energy use, kt); Emissions Total EMT-ENO-ASL (emissions N2O-all sectors with LULUCF).
Applsci 16 03236 g006
Applsci 16 03236 g007
Figure 7. Relationships between the observed values and predicted ones, for bioenergy consumption, in 2023.
Figure 7. Relationships between the observed values and predicted ones, for bioenergy consumption, in 2023.
Applsci 16 03236 g007
Applsci 16 03236 g008
Figure 8. The most important predictors for bioenergy consumption in 2023. Note: Emissions Total EMT-ECH-AGS (emissions CH4-agrifood systems, kt); Bioenergy BIO-PRO-TOB (total bioenergy production, Tj); Bioenergy BIO-PRO-Biogasoline (bioenergy production-biogasoline, Tj); Bioenergy BIO-PRO-SOB (bioenergy production-solid biofuels, Tj); Bioenergy BIO-PRO-OTV (bioenergy production-other vegetal material and residues, Tj); Bioenergy BIO-PRO-Biod (bioenergy production-biodiesel, Tj); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Emissions Total EMT-ECH-ENF (emissions CH4-enteric fermentation, kt); Food Balance FOB-DSQ-RIaP (domestic supply quantity-rice and products, 1000 t); Bioenergy BIO-PRO-Fuelw (bioenergy production-fuelwood, Tj).
Figure 8. The most important predictors for bioenergy consumption in 2023. Note: Emissions Total EMT-ECH-AGS (emissions CH4-agrifood systems, kt); Bioenergy BIO-PRO-TOB (total bioenergy production, Tj); Bioenergy BIO-PRO-Biogasoline (bioenergy production-biogasoline, Tj); Bioenergy BIO-PRO-SOB (bioenergy production-solid biofuels, Tj); Bioenergy BIO-PRO-OTV (bioenergy production-other vegetal material and residues, Tj); Bioenergy BIO-PRO-Biod (bioenergy production-biodiesel, Tj); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Emissions Total EMT-ECH-ENF (emissions CH4-enteric fermentation, kt); Food Balance FOB-DSQ-RIaP (domestic supply quantity-rice and products, 1000 t); Bioenergy BIO-PRO-Fuelw (bioenergy production-fuelwood, Tj).
Applsci 16 03236 g008
Applsci 16 03236 g009
Figure 9. Relationships between the observed values and predicted ones, for bioenergy consumption (without inputs for different types of bioenergy production and bioenergy consumption), in 2023.
Figure 9. Relationships between the observed values and predicted ones, for bioenergy consumption (without inputs for different types of bioenergy production and bioenergy consumption), in 2023.
Applsci 16 03236 g009
Applsci 16 03236 g010
Figure 10. The most important predictors for bioenergy consumption (without inputs for different types of bioenergy production and bioenergy consumption), in 2023. Note: Crop Emissions CRO-BBD-ALC (burning crop residues-biomass burned dry matter-all crops, t); Emissions Total EMT-ECH-AGS (emissions CH4-agrifood systems, kt); Forestry FOR-PRO-WOFU (production-wood fuel, m3); Emissions Total EMT-ECH-ENF (emissions CH4-enteric fermentation, kt); Emissions Total EMT-ECH-MAM (emissions CH4, manure management, kt); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Population POP-RUP-POP (rural population, 1000 persons); Crop Emission CRO-ARH-Rice (area harvest-rice, ha); Crop Emissions CRO-BBD-Maiz (burning crop residues-biomass burned dry matter-maize, t); Food Balance FOB-DSQ-ALBE (domestic supply quantity-alcoholic beverages, 1000 t).
Figure 10. The most important predictors for bioenergy consumption (without inputs for different types of bioenergy production and bioenergy consumption), in 2023. Note: Crop Emissions CRO-BBD-ALC (burning crop residues-biomass burned dry matter-all crops, t); Emissions Total EMT-ECH-AGS (emissions CH4-agrifood systems, kt); Forestry FOR-PRO-WOFU (production-wood fuel, m3); Emissions Total EMT-ECH-ENF (emissions CH4-enteric fermentation, kt); Emissions Total EMT-ECH-MAM (emissions CH4, manure management, kt); Crop Emissions CRO-BEN-Maiz (burning crop residues-emissions N2O-maize, kt); Population POP-RUP-POP (rural population, 1000 persons); Crop Emission CRO-ARH-Rice (area harvest-rice, ha); Crop Emissions CRO-BBD-Maiz (burning crop residues-biomass burned dry matter-maize, t); Food Balance FOB-DSQ-ALBE (domestic supply quantity-alcoholic beverages, 1000 t).
Applsci 16 03236 g010
Table 1. Top countries with the highest total bioenergy production (Tj), in 2023.
Table 1. Top countries with the highest total bioenergy production (Tj), in 2023.
CountriesEnergy Production (Total Bioenergy, Tj)Percentage of the Total (%)
India8,605,12517.332
China5,228,71810.531
United States of America4,102,6868.263
Brazil4,094,5858.247
Ethiopia1,737,9913.501
Democratic Republic of Congo1,361,8262.743
Indonesia1,220,6612.459
Germany1,080,1252.175
Nigeria1,058,7642.132
Uganda916,7601.846
Thailand851,1781.714
Tanzania840,0601.692
Kenya758,2561.527
Pakistan695,4261.401
Bangladesh617,4831.244
France570,9801.150
Canada522,6291.053
Guatemala471,7060.950
Myanmar460,8340.928
Sweden454,5010.915
Nepal431,0500.868
Italy428,5300.863
Poland422,4330.851
Vietnam393,7270.793
Finland388,9920.783
Total world49,649,778100.000
Table 2. World bioenergy production (Tj), in 2023, by type of energy.
Table 2. World bioenergy production (Tj), in 2023, by type of energy.
ItemEnergy Production (Tj)Percentage of the Total (%)
Fuelwood29,548,26559.513
Other vegetal material and residues7,880,78515.873
Bagasse4,321,6208.704
Biogasoline2,400,2014.834
Biodiesel2,243,5284.519
Charcoal1,976,5573.981
Black liquor1,787,0183.599
Biogases1,081,3902.178
Animal waste347,3300.700
Other liquid biofuels30,8100.062
Bio jet kerosene88310.018
Total Bioenergy49,649,778100.000
Table 3. Top countries with the highest total bioenergy consumption (Tj), in 2023.
Table 3. Top countries with the highest total bioenergy consumption (Tj), in 2023.
CountriesEnergy Consumption (Total Bioenergy, Tj)Percentage of the Total (%)
India7,612,47319.285
China4,664,90611.818
United States of America3,333,8048.446
Brazil2,942,0677.453
Ethiopia1,378,6883.493
Indonesia1,228,6763.113
Democratic Republic of the Congo984,8032.495
Nigeria919,7832.330
Uganda793,3062.010
United Republic of Tanzania730,2011.850
Pakistan672,8231.705
Germany635,7951.611
Bangladesh608,5961.542
France471,4051.194
Myanmar452,1451.145
Kenya436,7061.106
Nepal430,3471.090
Canada417,3311.057
Thailand396,2711.004
Guatemala383,1430.971
Poland350,2970.887
Italy342,0650.867
Zimbabwe335,1110.849
Philippines325,0620.824
Mexico297,4210.753
Total world39,473,209100.000
Table 4. World bioenergy consumption (Tj), in 2023, by type of energy.
Table 4. World bioenergy consumption (Tj), in 2023, by type of energy.
ItemEnergy Consumption (Tj)Percentage of the Total (%)
Fuelwood22,335,64356.584
Other vegetal material and residues6,623,06716.779
Biogasoline2,337,3485.921
Bagasse2,232,6865.656
Biodiesel2,005,3275.080
Charcoal1,969,8754.990
Black liquor1,502,1883.806
Animal waste340,9560.864
Biogases108,1100.274
Other liquid biofuels17,3530.044
Bio jet kerosene6560.002
Total Bioenergy39,473,209100.000
Table 5. Results for the most accurate models (bioenergy production, in 2023 and testing set).
Table 5. Results for the most accurate models (bioenergy production, in 2023 and testing set).
ModelBuild TimeCorrelationNumber of FieldsRelative Error
XGBoost Linear50.8644430.379
CHAID50.594340.609
Tree-AS50.44821.226
Linear50.639671.471
XGBoost Tree50.8364431.583
Note: XGBoost Linear (gradient boosting algorithm with a linear model); CHAID (Chi-squared Automatic Interaction Detection); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 6. Results for the most accurate models (bioenergy production, in 2023 and training set).
Table 6. Results for the most accurate models (bioenergy production, in 2023 and training set).
ModelBuild TimeCorrelationNumber of FieldsRelative Error
XGBoost Linear50.9994430.002
Linear50.997670.006
XGBoost Tree50.9964430.021
CHAID50.956340.087
Tree-AS50.40120.839
Note: XGBoost Linear (gradient boosting algorithm with a linear model); CHAID (Chi-squared Automatic Interaction Detection); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 7. Results for the most accurate models (bioenergy production, in 2023 and testing set), without inputs for different types of bioenergy production and bioenergy consumption.
Table 7. Results for the most accurate models (bioenergy production, in 2023 and testing set), without inputs for different types of bioenergy production and bioenergy consumption.
ModelBuild TimeCorrelationNumber of FieldsRelative Error
XGBoost Tree60.8754290.309
CHAID60.477180.843
Tree-AS60.44821.226
Linear60.593711.490
XGBoost Linear60.5124292.840
Note: XGBoost Linear (gradient boosting algorithm with a linear model); CHAID (Chi-squared Automatic Interaction Detection); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 8. Results for the most accurate models (bioenergy production, in 2023 and training set), without inputs for different types of bioenergy production and bioenergy consumption.
Table 8. Results for the most accurate models (bioenergy production, in 2023 and training set), without inputs for different types of bioenergy production and bioenergy consumption.
ModelBuild TimeCorrelationNumber of FieldsRelative Error
Linear60.997710.006
XGBoost Linear60.9964290.007
XGBoost Tree60.9964290.021
CHAID60.848180.281
Tree-AS60.40120.840
Note: XGBoost Linear (gradient boosting algorithm with a linear model); CHAID (Chi-squared Automatic Interaction Detection); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 9. Summary statistics (bioenergy production, in 2023).
Table 9. Summary statistics (bioenergy production, in 2023).
VariableObservationsMeanStandard DeviationMinMax
lnBioenergyBIO-PRO-TOB1999.6153.421−1.47715.968
lnCropEmissionsCRO-BBD-ALC18011.9703.216−0.43118.073
lnForestryFOR-PRO-WOFU19813.6543.0653.95119.512
lnFoodBalanceFOB-DSQ-COTOI552.2391.9780.0007.195
Table 10. Linear regression results with cross-section approach (bioenergy production, in 2023).
Table 10. Linear regression results with cross-section approach (bioenergy production, in 2023).
lnBioenergyBIO-PRO-TOBCoefficientRobust Standard ErrortP > t[95% Conf. Interval]
lnCropEmissionsCRO-BBD-ALC0.2250.2370.9500.348−0.2520.702
lnForestryFOR-PRO-WOFU0.7480.1554.8200.0000.4361.060
lnFoodBalanceFOB-DSQ-COTOI0.0370.1180.3200.753−0.1990.274
_cons−3.0582.279−1.3400.186−7.6351.518
VIF2.080
Breusch–Pagan/Cook–Weisberg test for heteroskedasticity44.780 (0.000)
Ramsey RESET test for omitted variables0.860 (0.466)
Table 11. Results for the most accurate models (bioenergy consumption, in 2023 and testing set).
Table 11. Results for the most accurate models (bioenergy consumption, in 2023 and testing set).
ModelBuild TimeCorrelationNumber of FieldsRelative Error
XGBoost Tree50.9424430.230
XGBoost Linear50.8524430.378
Tree-AS50.43821.100
Linear50.643911.696
C&R Tree50.559312.320
Note: XGBoost Linear (gradient boosting algorithm with a linear model); C&R Tree (Classification and Regression Tree); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 12. Results for the most accurate models (bioenergy consumption, in 2023 and training set).
Table 12. Results for the most accurate models (bioenergy consumption, in 2023 and training set).
ModelBuild TimeCorrelationNumber of FieldsRelative Error
XGBoost Linear50.9994430.002
Linear50.999910.002
XGBoost Tree50.9964430.022
C&R Tree50.940310.117
Tree-AS50.39420.845
Note: XGBoost Linear (gradient boosting algorithm with a linear model); C&R Tree (Classification and Regression Tree); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 13. Results for the most accurate models (bioenergy consumption, in 2023 and testing set), without inputs for different types of bioenergy production and bioenergy consumption.
Table 13. Results for the most accurate models (bioenergy consumption, in 2023 and testing set), without inputs for different types of bioenergy production and bioenergy consumption.
ModelBuild TimeCorrelationNumber of FieldsRelative Error
XGBoost Tree60.8234290.376
Tree-AS60.43821.100
Linear60.483751.962
C&R Tree60.656362.296
XGBoost Linear60.5004293.179
Note: XGBoost Linear (gradient boosting algorithm with a linear model); C&R Tree (Classification and Regression Tree); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 14. Results for the most accurate models (bioenergy consumption, in 2023 and training set), without inputs for different types of bioenergy production and bioenergy consumption.
Table 14. Results for the most accurate models (bioenergy consumption, in 2023 and training set), without inputs for different types of bioenergy production and bioenergy consumption.
ModelBuild TimeCorrelationNumber of FieldsRelative Error
Linear60.997750.005
XGBoost Linear60.9974290.007
XGBoost Tree60.9964290.022
C&R Tree60.935360.126
Tree-AS60.39420.845
Note: XGBoost Linear (gradient boosting algorithm with a linear model); C&R Tree (Classification and Regression Tree); Tree-AS (decision trees); Linear (linear regression); XGBoost Tree (gradient boosting algorithm with a tree model).
Table 15. Summary statistics (bioenergy consumption, in 2023).
Table 15. Summary statistics (bioenergy consumption, in 2023).
VariableObservationsMeanStandard DeviationMinMax
lnBioenergyBIO-COM-TOB2148.8663.803−2.56015.845
lnCropEmissionsCRO-BBD-ALC18011.9703.216−0.43118.073
lnEmissionsTotalEMT-ECH-AGS2273.9053.623−9.21010.207
lnForestryFOR-PRO-WOFU19813.6543.0653.95119.512
lnEmissionTotalEMT-ECH-ENF1963.9172.993−6.0759.586
Table 16. Linear regression results with cross-section approach (bioenergy consumption, in 2023).
Table 16. Linear regression results with cross-section approach (bioenergy consumption, in 2023).
lnBioenergyBIO-COM-TOBCoefficientRobust Standard ErrortP > t[95% Conf. Interval]Coefficient
lnCropEmissionsCRO-BBD-ALC0.1150.0552.1000.0380.0070.224
lnEmissionsTotalEMT-ECH-AGS0.1860.0712.6300.0090.0460.326
lnForestryFOR-PRO-WOFU0.6800.0709.6900.0000.5420.819
_cons−1.9590.612−3.2000.002−3.167−0.752
VIF3.420
Breusch–Pagan/Cook–Weisberg test for heteroskedasticity9.260 (0.002)
Ramsey RESET test for omitted variables1.330 (0.268)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Martinho, V.J.P.D. Bioenergy Production and Consumption Prediction: The Best Predictors for the Best Machine Learning Models from Hundreds of Variables. Appl. Sci. 2026, 16, 3236. https://doi.org/10.3390/app16073236

AMA Style

Martinho VJPD. Bioenergy Production and Consumption Prediction: The Best Predictors for the Best Machine Learning Models from Hundreds of Variables. Applied Sciences. 2026; 16(7):3236. https://doi.org/10.3390/app16073236

Chicago/Turabian Style

Martinho, Vítor João Pereira Domingues. 2026. "Bioenergy Production and Consumption Prediction: The Best Predictors for the Best Machine Learning Models from Hundreds of Variables" Applied Sciences 16, no. 7: 3236. https://doi.org/10.3390/app16073236

APA Style

Martinho, V. J. P. D. (2026). Bioenergy Production and Consumption Prediction: The Best Predictors for the Best Machine Learning Models from Hundreds of Variables. Applied Sciences, 16(7), 3236. https://doi.org/10.3390/app16073236

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.
Back to TopTop