Next Article in Journal
Risk of Heart Failure in Patients with ST-Elevation Myocardial Infarction Receiving Drug-Eluting Stent Implantation and Undefined Duration of Antiplatelets
Next Article in Special Issue
Women’s Experiences with Compliance with Pelvic Floor Home Exercise Therapy and Lifestyle Changes for Pelvic Organ Prolapse Symptoms: A Qualitative Study
Previous Article in Journal
Basic Research in Atherosclerosis: Technologies of Personalized Medicine
Previous Article in Special Issue
What Influences Women to Adhere to Pelvic Floor Exercises after Physiotherapy Treatment? A Qualitative Study for Individualized Pelvic Health Care
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:

Extracting New Temporal Features to Improve the Interpretability of Undiagnosed Type 2 Diabetes Mellitus Prediction Models

Institute of Informatics, Faculty of Electrical Engineering and Computer Science, University of Maribor, 2000 Maribor, Slovenia
Faculty of Health Sciences, University of Maribor, 2000 Maribor, Slovenia
Usher Institute, University of Edinburgh, Edinburgh EH8 9YL, UK
Author to whom correspondence should be addressed.
J. Pers. Med. 2022, 12(3), 368;
Received: 6 December 2021 / Revised: 13 February 2022 / Accepted: 25 February 2022 / Published: 28 February 2022
(This article belongs to the Special Issue Personalized Nursing and Health Care)


Type 2 diabetes mellitus (T2DM) often results in high morbidity and mortality. In addition, T2DM presents a substantial financial burden for individuals and their families, health systems, and societies. According to studies and reports, globally, the incidence and prevalence of T2DM are increasing rapidly. Several models have been built to predict T2DM onset in the future or detect undiagnosed T2DM in patients. Additional to the performance of such models, their interpretability is crucial for health experts, especially in personalized clinical prediction models. Data collected over 42 months from health check-up examinations and prescribed drugs data repositories of four primary healthcare providers were used in this study. We propose a framework consisting of LogicRegression based feature extraction and Least Absolute Shrinkage and Selection operator based prediction modeling for undiagnosed T2DM prediction. Performance of the models was measured using Area under the ROC curve (AUC) with corresponding confidence intervals. Results show that using LogicRegression based feature extraction resulted in simpler models, which are easier for healthcare experts to interpret, especially in cases with many binary features. Models developed using the proposed framework resulted in an AUC of 0.818 (95% Confidence Interval (CI): 0.812−0.823) that was comparable to more complex models (i.e., models with a larger number of features), where all features were included in prediction model development with the AUC of 0.816 (95% CI: 0.810−0.822). However, the difference in the number of used features was significant. This study proposes a framework for building interpretable models in healthcare that can contribute to higher trust in prediction models from healthcare experts.

Graphical Abstract

1. Introduction

Morbidity and mortality are often results of Type 2 diabetes mellitus (T2DM). In addition, T2DM presents a substantial financial drain for individuals and families, health systems, and societies. Globally, the incidence and prevalence of T2DM are increasing rapidly [1]. In 2017, it was estimated that 425 million people had any diabetes (approx. 5.5% of the worldwide population), of which 90% had T2DM. According to projection estimations, the prevalence is going to increase substantially in the coming years; by 2045, for example, a 48% increase of prevalence from the above numbers is expected, or in absolute numbers, an estimated 629 million people (approx. 6.6% of the worldwide population) are expected to be suffering from any diabetes [2]. T2DM can also lead to a substantially increased risk of macrovascular and microvascular disease, especially in inadequate glycemic control [3]. Impaired fasting glucose typically leads to slow progression of T2DM and, more importantly, its symptoms may remain undetected for many years.
Electronic Health Records (EHR) enable researchers to perform predictive modeling by providing a large amount of data [4] and many links have been found between patient health, the environment, and clinical decisions [5]. Nowadays, data mining techniques are applied to various fields of science, including healthcare and medicine [6]. Usually, techniques such as pattern recognition, disease prediction, and classification are used. Although multiple methods are available to build prediction models, prediction accuracy and data validity are often not realistic for model application in practice. Models usually perform well in specific datasets used to build the prediction models but are frequently not adapted sufficiently well when used on other datasets [7].
There is growing interest in clinical prediction, but models’ interpretation is rarely based on end-user needs [8], and there is a lack of model interpretability techniques [9]. Interpretability of results based on predictive models is crucial in critical areas such as healthcare and is essential for adopting models. People often do not understand predictive models and therefore do not trust them [10]. LogicRegression can be used to improve the interpretability of predictive models.
LogicRegression is an adaptive classification and regression procedure which searches for Boolean (logic) combinations of binary variables that best explain the variability in the outcome [11,12]. LogicRegression looks for logical combinations of binary features. We can explain the variability of the outcome feature and thus reveal the features and interactions related to the response and whether they have predictive capabilities [11].
The purpose of this paper is to use LogicRegression to make final models less complex (i.e., with less features) and the features that appear in the interpretation of predictive models much more understandable. This is also important for health professionals, as they do not have the necessary knowledge to apply prediction models or interpret the results obtained. This is also important from the patient’s point of view and the provision of personalized healthcare. Simple interpretation will make it easier for the patient to understand the operation of the predictive model and outcome. The paper presents an example of using extracted features using Logic Regression to improve the personalized interpretability of the prediction models to the end-users.

2. Materials and Methods

2.1. Data

EHR data consisted of health check-ups and prescribed drugs data from four Slovenian primary healthcare providers for a period of approximately 3.5 years from 12 December 2014 to 27 July 2018. Data for 21,138 medical records and 114 potential useful features were exported from the healthcare information systems after the on-site anonymization process. Our first step was the removal of features with more than 20% of missing data (73 potential features remain). Since our focus when building prediction models was on the fasting plasma glucose level (FPGL) measurement (mmol/L) and results of Finnish Diabetes Risk Score (FINDRISC) features, which included Age, Gender, BMI, Waist circumference, Active_30_min, Medication, High_BS, Grocer, and Diab_fam we selected cases with all those values present (4086 such cases remained). We next removed (a) cases with more than 50% of the features were not available (4067 cases left), (b) removed all duplicate entries (in cases of multiple patient visits only the most recent visit was included) (3535 cases left), (c) cases not having a previous diabetes diagnosis (3176 cases left) and entries where: (d) FPGL was not reported giving us a total of 3120 records of patient visits were left for development of a prediction model to estimate the risk of undiagnosed T2DM. Data included demographics, questionnaire answers for lifestyle choices, physiological measurements, and prescribed medications for two time periods.
Binary features were created for prescribed drugs and questionnaire responses, which resulted in nine numeric and 161 binary features where specific drug related feature was coded as positive in cases where a patient was prescribed with the specific drug during the last 4 months prior to the visit. The target feature was binary, where positive cases were defined as having FPGL higher than 6.1 mmol/L consisting of 24.71% (n = 771) of patient visits.
We imputed the remaining missing values using the MissForest based approach [7], which on average meant features with 12.25% missing values as we initially already removed features with 20% or more missing values. MissForest is used to impute missing values particularly in the case of mixed-type data. It can be used to impute continuous and/or categorical data including complex interactions and nonlinear relations. The summary information of the basic predictive and target features can be seen in Table 1. Please see Table A1 for list of all features used in the experiments.

2.2. Experimental Setup

The data were split into 80% to derive five extracted features using Logic Regression [13] and 20% to build and evaluate the final prediction models.
Finally, we created three datasets with the following features: all numeric and binary (170), all numeric and logic (14), and all numeric, binary, and logic features (175). On each dataset, we built a predictive model separately using the same training data.
The Least Absolute Shrinkage and Selection Operator (LASSO) [13] was used to build prediction models. We repeated each 10-fold cross-validation ten times to estimate the variance in Area under the ROC curve (AUC) that was used as our classification performance metric.

3. Results

We split the results in this section into two parts. First, we present selected logic attributes extracted from the dataset for the undiagnosed T2DM prediction use case. Next, we present the performance evaluation of the model.

3.1. Feature Extraction Using LogicRegression Approach

To demonstrate the practical example of using LogicRegression based extraction of new features to improve interpretability of the prediction models, we provide the results of the first cross-validation run.
The selected use case resulted in five logic features (Table 2) extracted from the complete set of features.
In Table 3, we list all features that were selected in at least 50% of runs in our experiments with LASSO on the dataset with numeric and logic features, while Table 4 lists features for the dataset with numeric, binary and logic features. Frequency (freq) shows in how many experiment runs each feature appeared in the final set of features.
It can be observed that L1, L2, and L3 were used by prediction models derived from the data in all folds of all evaluation runs. Thus, confirming a high contribution of extracted logic features.
In the case of results from a much wider set of features (Table 4), we can see a higher variance in selection by the final prediction models. Four (L2, L3, L4, L5) logic features can be found among the varaibles that were selected in at least 50% of evaluation runs.

3.2. Performance Evaluation

In Figure 1, we summarize AUC and a selected number of features for all three datasets: no_logic (numeric and binary features), all_logic (numeric, binary, and logic features), and num_logic (numeric and logic features).
We can observe a slowly increasing average AUC from 0.816 (Standard Deviation (SD)) = 0.03) in no_logic to 0.819 (SD = 0.03) in all_logic and finally 0.829 (SD = 0.03) in the num_logic dataset. When looking at the number of selected feature averages and its variation, we can observe that it slowly increases from 21.7 (SD = 11.18) in no _logic to 23.7 (SD = 10.09) in all_logic but it then almost halves to 13.35 (SD = 0.63) in the num_logic dataset. The SD is steadily increasing in the first two cases, but then it decreases sharply to below 1 (SD = 0.63), which means that out of the 100 repetitions in 92 cases 13 or 14 features were selected in the num_logic dataset. This indicates a very stable final prediction models when comparing num_logic based solutions to no_logic or all_logic.

4. Discussion and Conclusions

In this paper, we compared three dimensionality reduction approaches to improve the interpretability of undiagnosed T2DM prediction models (Please note that the calibration of a prediction models was not the scope of this paper and presents a limitation). A simple LASSO regression approach is compared to two variants where a pre-selection of predictive features is conducted on the training set using LogicRegression to consequently simplify a final set of features obtained by the LASSO regression. We kept all original features with added logic features in the first variant, while in the second variant, we kept only numeric and logic features.
Results showed that logic features resulted in simpler models with lower number of features, which are potentially easier to interpret by healthcare experts. This is especially important in the field of personalized medicine. Measured AUC was similar to more complex models, where all features were included. It should be noted that although our method resulted in a lower number of features, some of the logic features may not be straightforward to interpret (e.g., the feature L3 in this paper). To address this issue, we plan to include an interactive system in our future work, where the user would specify the maximum number of original features included in generated logic features in cases where the final model would include many complex logic features. As a result of the current work, in cases when some of the final features are hard to interpret, we recommend that the user uses LogicRegression settings to adjust the complexity of final logic features for achieving satisfactory results.
When healthcare professionals and patients know which features are important in obtaining the outcome of a prediction model and how they can be combined, it helps to understand and increase the level of trust in the decision-making systems [10]. With greater interpretability of the model, we better understand and interpret the forecast for end-users and improve the support in decision-making for health professionals based on data [14]. More complex models such as deep neural networks [15] allow high accuracy but are difficult to explain. Simple models (e.g., decision trees) are less accurate but allow for more straightforward explanations [16]. Therefore, sophisticated machine learning models usually offer better performance than traditional simple models but are difficult for health professionals to understand. However, in many cases simple models also provide good classification performance, which is not significantly different from more complex models [17]. Our results confirm this hypothesis. Comprehensible models are known for their contribution to higher trust in prediction models from the end-users in healthcare.
Interpretability techniques are often categorized according to the time period used to develop the machine learning model [14]. Pre-model approaches are independent of the model and may be employed prior to making a choice on which model to use. Our approach presented in this study belongs in this group of interpretability approaches along with techniques such as Principal Component Analysis (PCA), t-Distributed Stochastic Neighbor Embedding (t-SNE), and some clustering techniques. While Molnar [18] classifies PCA, t-SNE, and clustering methods as interpretable methods, it is worth noting that the interpretability of attributes transformed using PCA, embeddings, or clusters cannot provide comprehensible medical interpretation, but can be used to visualize the results and highlight patterns of interest from an interpretability standpoint. The proposed approach is much more interpretable, despite the possible complex combinations of features that might occur as a result of LogicRegression.
During the experiments, we also observed the unstable behavior of logic regression, where different logic features were selected with each run of the cross-validation. Although this did not influence the average number of selected features it resulted in instability of the interpretability of the model. Another limitation are the combinations of the features used in extracted features. For example, the first extracted feature (L1) suggested that checking whether a person did not experience elevated blood sugar in the past should be accompanied by checking for sulfadiazine and trimethoprim use in the last 4 months – this extracted feature works as a protective factor as seen from Table 3. We see this as a disadvantage of logic regression since different conclusions can be made based on selected features. This could be resolved to some extent by using exhaustive search methods to extract logic features resulting in extremely long running times, presenting another drawback, especially in cases where personalized models would be built. To personalize the solution even further, it would be worth exploring the prediction model development for each specific patient at the time of the examination using the subset of the data where patients similar to the examined patient would be assigned a higher weight in comparison to other patients (boosting principle).
Although our work is the field of healthcare, we believe that our results can also be applied in other emerging fields of applied prediction modeling where interpretability of results is important such as security [19] or ecology [20]. In future work, we will explore effectiveness of our methods in the broader field of security, specifically, to help us understand how misinformation (e.g., intentionally misleading information) is being spread.

Author Contributions

Conceptualization, S.K., P.K. and G.Š.; methodology, S.K., P.K. and G.Š.; software, S.K., P.K. and G.Š.; validation, S.K., P.K. and G.Š.; formal analysis, S.K. and P.K.; investigation, S.K., P.K., L.G. and G.Š.; resources, P.K. and G.Š.; data curation, S.K. and P.K.; writing—original draft preparation, S.K., P.K., L.G. and G.Š.; writing—review and editing, S.K., P.K., L.G., N.F. and G.Š.; visualization, P.K.; supervision, G.Š.; project administration, G.Š.; funding acquisition, G.Š. All authors have read and agreed to the published version of the manuscript.


This research was funded by Slovenian Research Agency (grants number ARRS N2-0101 and ARRS P2-0057) and the European Union’s Horizon 2020 Research and Innovation Program under the Cybersecurity CONCORDIA project (GA No. 830927).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Restrictions apply to the availability of these data. Data was obtained from Nova vizija d.d. and are available from the authors with the permission of Nova vizija d.d.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Table A1. List of original features with description and their possible values. Please note that Nominal features were processed in such a way that for each possible value a new feature was generated. For example, the feature Q43 resulted in three features for each possible value (new features were named Q431, Q432 and Q433). Drug features are marked with the Anatomical Therapeutic Chemical (ATC) classification. The final set contained 170 features.
Table A1. List of original features with description and their possible values. Please note that Nominal features were processed in such a way that for each possible value a new feature was generated. For example, the feature Q43 resulted in three features for each possible value (new features were named Q431, Q432 and Q433). Drug features are marked with the Anatomical Therapeutic Chemical (ATC) classification. The final set contained 170 features.
AgeAge of the patientNumeric
GenderGender of the patientMale, Female
BMIBody Mass Index of the patientNumeric
Blood_pressureBlood pressure of the patientNumeric
WCWaist circumference of the patientNumeric
Heart_beatHeart beat of the patientNumeric
Body_weightBody weight of the patientNumeric
Body_heightBody height of the patientNumeric
Smoking_statusSmoking status of the patientNon-smoker, Smoker, Ex-smoker, Passive smoker
Eating_habitsAssessment of eating habitsAdequate, Satisfactory, Inadequate
Drinking_statusDrinking statusAbstinent, Less risky drinking, Risky, Harmful, Addictive
SDHSocial determinants of healthNot threatened, Medium threatened, Threatened
PASPhysical activity statusSufficient, Borderline, Insufficient
StressLevel of stressNot threatened, Threatened
RDRisk of depressionNo significant risk of depression, Risk of depression
Q18How often do you usually eat vegetables?Never Points, 4-6 times a week, 1x a day, More than 1x a day
Q16How many meals do you eat on average per day?2 or less, 3 to 5, 6 or more
Q2Are you physically active for at least 30 min/day?Yes, No
Q3Do you take medication to lower your blood pressure?Yes, No
Q30Do you have a habit of salting dishes at the table?Yes, No
Q32On average, which type of fat do you use most in food preparation or as a spread?Vegetable oils, Cream, Butter, Lard, Hard margarines, Soft margarines, High-fat spreads, Low-fat spreads, Chocolate spread, Peanut butter, Pate, Cream Spread, Mayonnaise
Q4Have you ever had your blood sugar measured?Yes, No
Q43How many times in a typical week do you engage in vigorous physical activity for at least 25 minutes each time to the point where you are breathing and sweating?0 or 1 times per week, 2 times per week, 3 or more times per week
Q44How many times in a typical week do you engage in moderate physical activity for at least 30 minutes each time, to the extent that you breathe a little faster and warm up?0 or 1 times per week, 2 to 4 times per week, 5 or more times per week
Q47How often have you drunk drinks containing alcohol in the last 12 months?Never, Once a month or less, 2 to 4 times a month, 2 to 3 times a week, 4 or more times a week
Q48In the last 12 months, how many measures of a drink containing alcohol did you usually have when you were drinking?Zero to 1 measure, 2 measures, 3 or 4 measures, 5 or 6 measures, 7 or more
Q49In the last 12 months, how often have you had 6 or more sips on one occasion for men and 4 or more sips on one occasion for women?Never, Less than once a month, 1 to 3 times a month, 1 to 3 times a week, Daily or almost daily
Q51In the last 12 months, how often have you needed an alcoholic drink in the morning to recover from excessive drinking the day before?Never, Less than once a month, 1 to 3 times a month, 1 to 3 times a week, Daily or almost daily
Q57How often do you feel tense, stressed or under a lot of pressure?Never, Rarely, Occasionally, Often, Every day
Q58How do you manage the tensions, stresses and pressures you experience in your life?Easily, Able to, Able to with more efforts, Very difficult, Can’t
Q59How often in the past 2 weeks have you felt little interest and satisfaction in the things you do?Not at all, A few days, More than half the days, Almost every day
Q6Does family have diabetes?No, Outer family, Inner family
Q60How often have you felt depressed, depressed, despairing in the past 2 weeks?Not at all, A few days, More than half the days, Almost every day
Q69Please indicate the last school you attended.Primary school incomplete, Primary school, 2 or 3-year vocational school, 4-year secondary school or gymnasium, Graduate, Postgraduate
Q70What is your current employment status?Employed, Self-employed, Unemployed, Student, Retired, Disabled pensioner, Permanently disabled, Housewife
Q71How do you get through the month based in income?Good, Occasional problems, I have problems
ATC_A02BC01OmeprazoleBinary (0,1)
ATC_A02BC02PantoprazoleBinary (0,1)
ATC_A11CC05ColecalciferolBinary (0,1)
ATC_B01AC06Acetylsalicylic acid Binary (0,1)
ATC_C03BA11IndapamideBinary (0,1)
ATC_C07AB07BisoprololBinary (0,1)
ATC_C09AA04PerindoprilBinary (0,1)
ATC_D01AC01Clotrimazole Binary (0,1)
ATC_D01AE15Terbinafine Binary (0,1)
ATC_D07AC13Mometasone Binary (0,1)
ATC_G04BD09Trospium Binary (0,1)
ATC_J01CA04AmoxicillinBinary (0,1)
ATC_J01CE10Benzathine phenoxymethylpenicillin Binary (0,1)
ATC_J01CR02Amoxicillin and beta-lactamase inhibitor Binary (0,1)
ATC_J01EE01Sulfadiazine /trimethoprimBinary (0,1)
ATC_J01FA10Azithromycin Binary (0,1)
ATC_M01AB05DiclofenacBinary (0,1)
ATC_M01AE01Ibuprofen Binary (0,1)
ATC_M01AE02NaproxenBinary (0,1)
ATC_N02AJ13Tramadol and paracetamol Binary (0,1)
ATC_N02BB02Metamizole sodium Binary (0,1)
ATC_N02BE01ParacetamolBinary (0,1)
ATC_N05BA08Bromazepam Binary (0,1)
ATC_N05BA12AlprazolamBinary (0,1)
ATC_N05CF02Zolpidem Binary (0,1)
ATC_R01AD09Mometasone Binary (0,1)
ATC_R03AC02Salbutamol Binary (0,1)
ATC_R03AL01Fenoterol and ipratropium bromide Binary (0,1)
ATC_R06AE07Cetirizine Binary (0,1)
ATC_R06AX13LoratadineBinary (0,1)
ATC_S01AA12TobramycinBinary (0,1)


  1. Einarson, T.R.; Acs, A.; Ludwig, C.; Panton, U.H. Prevalence of cardiovascular disease in type 2 diabetes: A systematic literature review of scientific evidence from across the world in 2007–2017. Cardiovasc. Diabetol. 2018, 17, 83. [Google Scholar] [CrossRef][Green Version]
  2. International Diabetes Federation. IDF Diabetes Atlas 2021, 10th ed.; IDF: Brussels, Belgium, 2021. [Google Scholar]
  3. Mohammedi, K.; Woodward, M.; Marre, M.; Colagiuri, S.; Cooper, M.; Harrap, S.; Mancia, G.; Poulter, N.; Williams, B.; Zoungas, S.; et al. Comparative effects of microvascular and macrovascular disease on the risk of major outcomes in patients with type 2 diabetes. Cardiovasc. Diabetol. 2017, 16, 95. [Google Scholar] [CrossRef]
  4. Steele, A.J.; Denaxas, S.C.; Shah, A.D.; Hemingway, H.; Luscombe, N.M. Machine learning models in electronic health records can outperform conventional survival models for predicting patient mortality in coronary artery disease. PLoS ONE 2018, 13, e0202344. [Google Scholar] [CrossRef] [PubMed][Green Version]
  5. La Cava, W.; Bauer, C.; Moore, J.H.; Pendergrass, S.A. Interpretation of machine learning predictions for patient outcomes in electronic health records. AMIA Annu. Symp. Proc. 2019, 2019, 572–581. [Google Scholar]
  6. Birjandi, S.M.; Khasteh, S.H. A survey on data mining techniques used in medicine. J. Diabetes Metab. Disord. 2021, 20, 2055–2071. [Google Scholar] [CrossRef]
  7. Stekhoven, D.J.; Bühlmann, P. Missforest—Non-parametric missing value imputation for mixed-type data. Bioinformatics 2012, 28, 112–118. [Google Scholar] [CrossRef][Green Version]
  8. Barda, A.J.; Horvat, C.M.; Hochheiser, H. A qualitative research framework for the design of user-centered displays of explanations for machine learning model predictions in healthcare. BMC Med. Inform. Decis. Mak. 2020, 20, 257. [Google Scholar] [CrossRef]
  9. Elshawi, R.; Al-Mallah, M.H.; Sakr, S. On the interpretability of machine learning-based model for predicting hypertension. BMC Med. Inform. Decis. Mak. 2019, 19, 146. [Google Scholar] [CrossRef] [PubMed][Green Version]
  10. Lakkaraju, H.; Bach, S.H.; Leskovec, J. Interpretable decision sets: A joint framework for description and prediction. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; pp. 1675–1684. [Google Scholar]
  11. Schwender, H.; Ruczinski, I. Logic regression and its extensions. Adv. Genet. 2010, 72, 25–45. [Google Scholar] [PubMed]
  12. Ruczinski, I.; Kooperberg, C.; LeBlanc, M. Logic regression. J. Comput. Graph. Stat. 2003, 12, 475–511. [Google Scholar] [CrossRef]
  13. Friedman, J.; Hastie, T.; Tibshirani, R. Regularization paths for generalized linear models via coordinate descent. J. Stat. Softw. 2010, 33. [Google Scholar] [CrossRef][Green Version]
  14. Stiglic, G.; Kocbek, P.; Fijacko, N.; Zitnik, M.; Verbert, K.; Cilar, L. Interpretability of machine learning-based prediction models in healthcare. Wiley Interdiscip. Rev. Data Min. Knowl. Discov. 2020, 10, e1379. [Google Scholar] [CrossRef]
  15. Emmert-Streib, F.; Yang, Z.; Feng, H.; Tripathi, S.; Dehmer, M. An introductory review of deep learning for prediction models with big data. Front. Artif. Intell. 2020, 3, 4. [Google Scholar] [CrossRef] [PubMed][Green Version]
  16. Lim, T.S.; Loh, W.Y.; Shih, Y.S. A comparison of prediction accuracy, complexity, and training time of thirty-three old and new classification algorithms. Mach. Learn. 2000, 40, 203–228. [Google Scholar] [CrossRef]
  17. Stiglic, G.; Kocbek, S.; Pernek, I.; Kokol, P. Comprehensive decision tree models in bioinformatics. PLoS ONE 2012, 7, e33812. [Google Scholar] [CrossRef]
  18. Molnar, C. Interpretable Machine Learning; Research Triangle, NC, USA, 2020. [Google Scholar]
  19. Brigugilio, W.R. Machine Learning Interpretability in Malware Detection. Ph.D. Dissertation, University of Windsor, Windsor, ON, Canada, 2020. [Google Scholar]
  20. Lucas, T.C. A translucent box: Interpretable machine learning in ecology. Ecol. Monogr. 2020, 90, e01422. [Google Scholar] [CrossRef]
Figure 1. Selected features with the Least Absolute Shrinkage and Selection Operator (LASSO) on the dataset with numeric and logic features.
Figure 1. Selected features with the Least Absolute Shrinkage and Selection Operator (LASSO) on the dataset with numeric and logic features.
Jpm 12 00368 g001
Table 1. Summary table basic predictive and target features for healthcare centers.
Table 1. Summary table basic predictive and target features for healthcare centers.
Original Feature NameDescriptionFPGL ≤ 6.1 mmol/L
[75.29% [n = 2349]]
FPGL > 6.1 mmol/L
[24.71% [n = 771]]
Age [mean (standard deviation − SD)]Age in years56.07 (SD = 13.2)61.77 (SD = 10.98)
Gender_M [%(n)]Percentage of males37.16 (n = 873)54.47 (n = 420)
BMI [mean (SD)]Body mass index28.89 (SD = 5.39)32.16 (SD = 13.21)
WC [mean (SD)]Waist circumference in cm96.25 (SD = 13.89)103.48 (SD = 13.8)
Active_30_min (Q2) [%(n)]Active at least 30 minutes a day?64.88 (n = 1524)52.27 (n = 403)
Medication (Q3)
Blood pressure medication?40.19 (n = 944)60.18 (n = 464)
High_BS [%(n)] (Q4)Ever measured high blood sugar?7.32 (n = 172)47.47 (n = 366)
Grocer [%(n)] (Q18)Eat vegetable/fruit daily?90.59 (n = 2128)78.99 (n = 609)
Diab_fam [%(n)] (Q6)Diabetes in family?69.65 (n = 1636)61.74 (n = 476)
FPGL [mean (SD)]Fasting plasma glucose level5.26 (SD = 0.44)6.74 (SD = 0.8)
Table 2. Extracted logic features with corresponding LogicRegression rules and descriptions.
Table 2. Extracted logic features with corresponding LogicRegression rules and descriptions.
L1(ATC_J01EE01 or (not Q41))Prescribed sulfadiazine and trimethoprim, or never measured high blood sugar.
L2Q51Seldom eat fruit and vegetable.
L3((ATC_M01AE02 and ATC_J01CE10) or (not SE))Prescribed naproxen and benzathine phenoxymethylpenicillin or not socially endangered.
L4Q494Daily consumption of alcohol in the last 12 months.
L5(MSE or ATC_D01AE15)Medium socially endangered or prescribed antifungals for dermatological use.
Table 3. Selected features with the Least Absolute Shrinkage and Selection Operator (LASSO) on the dataset with numeric and logic features.
Table 3. Selected features with the Least Absolute Shrinkage and Selection Operator (LASSO) on the dataset with numeric and logic features.
+Blood_pressure100Blood pressure
+BMI100Body mass index
+WC100Waist circumference
−L1100Logic feature 1
−L2100Logic feature 2
−L3100Logic feature 3
−Body_height99Body height
+Body_weight83Body weight
Table 4. Selected features with LASSO on the dataset with binary, numeric, and logic features.
Table 4. Selected features with LASSO on the dataset with binary, numeric, and logic features.
−L3100Logic feature 3
−L4100Logic feature 4
+L5100Logic feature 5
+Blood_pressure100Blood pressure
+WC100Waist circumference in cm
+Age100Age in years
+Q45100Ever measured high blood sugar? Yes
+Q3293Using drug(s) for lowering blood pressure
+Body_weight87Body weight
+L279Logic attribute 2
−Q32178Most often used oil is vegetable oil
−Non_drinker77No alcohol consumption
−Q58375Handle stress with hardship
+Q6274Parent, brother, or sister have diabetes
+BMI69Body Mass Index
−Q161632 meals per day on average
−Q30151No habit of using salt at the table
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Kocbek, S.; Kocbek, P.; Gosak, L.; Fijačko, N.; Štiglic, G. Extracting New Temporal Features to Improve the Interpretability of Undiagnosed Type 2 Diabetes Mellitus Prediction Models. J. Pers. Med. 2022, 12, 368.

AMA Style

Kocbek S, Kocbek P, Gosak L, Fijačko N, Štiglic G. Extracting New Temporal Features to Improve the Interpretability of Undiagnosed Type 2 Diabetes Mellitus Prediction Models. Journal of Personalized Medicine. 2022; 12(3):368.

Chicago/Turabian Style

Kocbek, Simon, Primož Kocbek, Lucija Gosak, Nino Fijačko, and Gregor Štiglic. 2022. "Extracting New Temporal Features to Improve the Interpretability of Undiagnosed Type 2 Diabetes Mellitus Prediction Models" Journal of Personalized Medicine 12, no. 3: 368.

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

Article Metrics

Back to TopTop