Search Results (7)

Visualization techniques for understanding and explaining machine learning models have gained significant attention. One such technique is the decision map, which creates a 2D depiction of the decision behavior of classifiers trained on high-dimensional data. While several decision map techniques have been proposed recently, such as Decision Boundary Maps (DBMs), Supervised Decision Boundary Maps (SDBMs), and DeepView (DV), there is no framework for comprehensively evaluating and comparing these techniques. In this paper, we propose such a framework by combining quantitative metrics and qualitative assessment. We apply our framework to DBM, SDBM, and DV using a range of both synthetic and real-world classification techniques and datasets. Our results show that none of the evaluated decision-map techniques consistently outperforms the others in all measured aspects. Separately, our analysis exposes several previously unknown properties and limitations of decision-map techniques. To support practitioners, we also propose a workflow for selecting the most appropriate decision-map technique for given datasets, classifiers, and requirements of the application at hand. Full article

Sexual dimorphism exists widely in animals, manifesting in different forms, such as body size, color, shape, unique characteristics, behavior, and sound. Of these, body mass dimorphism is the most obvious. Studies of evolutionary and ontogenetic development and adaptation mechanisms of animals’ sexual dimorphism in body mass (SDBM), allow us to understand how environment, social group size, diet, and other external factors have driven the selection of sexual dimorphism. There are fewer reports of the ontogenetic development of sexual dimorphism in body mass in Rhinopithecus. This study explores the ontogenetic development pattern of SDBM in wild black-and-white snub-nosed monkeys (R. bieti), and the causes resulting in extreme sexual dimorphism compared to other colobines. A significant dimorphism with a ratio of 1.27 (p < 0.001) appears when females enter the reproductive period around six years old, reaching a peak (1.85, p < 0.001) when males become sexually mature. After the age of eight, the SDBM falls to 1.78, but is still significant (p < 0.001). The results also indicate that males had a longer body mass growth period than females (8 years vs. 5 years); females in larger breeding units had a significantly higher SDBM than those in smaller ones (2.12 vs. 1.93, p < 0.01). A comparative analysis with other colobines further clarifies that Rhinopithecus and Nasalis, which both have multilevel social organization, have the highest degree of SDBM among all colobines. The large SDBM in R. bieti can be explained through Bergman’s and Rensch’s rules. Overall, environmental adaptation, a distinctive alimentary system, and a complex social structure contribute to R. bieti having such a remarkable SDBM compared to other colobines. In addition, we found that females’ choice for males may not be significantly related to the development of SDBM. Full article

Machine learning provides solutions to a diverse range of problems in high-dimensional datasets in geosciences. However, machine learning is generally criticized for being an enigmatic black box as it focusses on results but ignores the processes. To address this issue, we used supervised decision boundary maps (SDBM) to visually illustrate and interpret the machine learning process. We constructed a SDBM to classify the ore genetics from 1551 trace element data of apatite in various types of deposits. Attribute-based visual explanation of multidimensional projections (A-MPs) was introduced to SDBM to further demonstrate the correlation between features and machine learning process. Our results show that SDBM explores the interpretability of machine learning process and the A-MPs approach reveals the role of trace elements in machine learning classification. Combining SDBM and A-MPs methods, we propose intuitive and accurate discrimination diagrams and the most indicative elements for ore genetic types. Our work provides novel insights for the visualization application of geo-machine learning, which is expected to be a powerful tool for high-dimensional geochemical data analysis and mineral deposit exploration. Full article

(1) Background: For the elderly and disabled, self-driving is very important for social participation. An understanding of changing driving conditions is essential in order to drive safely. This study aimed to develop a web-based Korean Mini-Driving Scene Screening Test (MDSST) and to verify its reliability and validity for clinical application. (2) Methods: We developed a web-based MDSST, and its content validity was verified by an expert group. The tests were conducted with 102 elderly drivers to verify the internal consistency and reliability of items, and the validity of convergence with the existing Korean-Safe Driving Behavior Measure (K-SDBM) and the Korean-Adelaide Driving Self-Efficacy Scale (K-ADSES) driving tests was also verified. The test–retest reliability was verified using 54 individuals who participated in the initial test. (3) Results: The average content validity index of MDSST was 0.90, and the average internal consistency of all items was 0.822, indicating high content validity and internal consistency. The exploratory factor analysis for construct validity, the KOM value of the data, was 0.658, and Bartlett’s sphericity test also showed a strongly significant result. The four factors were road traffic and signal perception, situation understanding, risk factor recognition, and situation prediction. The explanatory power was reliable at 61.27%. For the convergence validation, MDSST and K-SDBM showed r = 0.435 and K-ADSES showed r = 0.346, showing a moderate correlation. In the evaluation–reevaluation reliability verification, the reliability increased to r = 0.952. (4) Conclusions: The web-based MDSST test developed in this study is a useful tool for detecting and understanding real-world driving situations faced by elderly drivers. It is hoped that the MDSST test can be applied more widely as a driving ability test that can be used in the clinical field of driving rehabilitation. Full article

Trajectory data allow the study of the behavior of moving objects, from humans to animals. Wireless communication, mobile devices, and technologies such as Global Positioning System (GPS) have contributed to the growth of the trajectory research field. With the considerable growth in the volume of trajectory data, storing such data into Spatial Database Management Systems (SDBMS) has become challenging. Hence, Spatial Big Data emerges as a data management technology for indexing, storing, and retrieving large volumes of spatio-temporal data. A Data Warehouse (DW) is one of the premier Big Data analysis and complex query processing infrastructures. Trajectory Data Warehouses (TDW) emerge as a DW dedicated to trajectory data analysis. A list and discussions on problems that use TDW and forward directions for the works in this field are the primary goals of this survey. This article collected state-of-the-art on Big Data trajectory analytics. Understanding how the research in trajectory data are being conducted, what main techniques have been used, and how they can be embedded in an Online Analytical Processing (OLAP) architecture can enhance the efficiency and development of decision-making systems that deal with trajectory data. Full article

Over recent decades, more and more cities worldwide have created semantic 3D city models of their built environments based on standards across multiple domains. 3D city models, which are often employed for a large range of tasks, go far beyond pure visualization. Due to different spatial scale requirements for planning and managing various built environments, integration of Geographic Information Systems (GIS) and Building Information Modeling (BIM) has emerged in recent years. Focus is now shifting to Precinct Information Modeling (PIM) which is in a more general sense to built-environment modeling. As scales change so do options to perform information modeling for different applications. How to implement data interoperability across these digital representations, therefore, becomes an emerging challenge. Moreover, with the growth of multi-source heterogeneous data consisting of semantic and varying 2D/3D spatial representations, data management becomes feasible for facilitating the development and deployment of PIM applications. How to use heterogeneous data in an integrating manner to further express PIM is an open and comprehensive topic. In this paper, we develop a semantic PIM based on multi-source heterogeneous data. Then, we tackle spatial data management problems in a Spatial Database Management System (SDBMS) solution for our defined unified model. Case studies on the University of New South Wales (UNSW) campus demonstrate the efficiency of our solution. Full article

The exceptional potential of the graphene has not been yet fully translated into the Al matrix to achieve high-performance Al nanocomposite. This is due to some critical issues faced by graphene during its processing such as the dispersion uniformity, structure damage, compatibility/wettability, and low graphene embedding content in Al matrix. In the present work, a new integrative method was adopted and named as “solvent dispersion and ball milling” (SDBM) to address the issues above efficiently in a single approach. This strategy involves effective graphene nanoplatelets (GNPs) solvent dispersion via surfactant decoration and solution ball milling employed to polyvinyl alcohol (PVA) coated Al with various GNPs content (0.5, 1 and 1.5 wt. %). Flaky Al powder morphology attained by optimizing ball milling parameters and used for further processing with GNPs. Detailed powders characterizations were conducted to investigate morphology, graphene dispersion, group functionalities by FTIR (Fourier transform infrared spectroscopy) spectroscopy and crystallinity by powder XRD (X-ray diffraction)analysis. Compaction behaviour and spring back effect of the GNPs/Al powders was also investigated at different compaction pressure (300 to 600 Mpa) and varying GNPs fractions. In response, green and sintered relative density (%) along with effect on the hardness of the nanocomposites samples were examined. Conclusively, in comparison with the unreinforced Al, GNP/Al nanocomposite with 1.5 wt. % GNPs exhibited the highest hardness gives 62% maximum increase than pure Al validates the effectiveness of the approach produces high fraction uniformly dispersed GNPs in Al matrix. Full article