Multidimensional Classification Research Articles

Taste the music: Modality-general representation of affective states derived from auditory and gustatory stimuli

Prior studies have extensively examined modality-general representation of affect across various sensory modalities, particularly focusing on auditory and visual stimuli. However, little research has explored the modality-general representation of affect between gustatory and other sensory modalities. This study aimed to investigate whether the affective responses induced by tastes and musical pieces could be predicted within and across modalities. For each modality, eight stimuli were chosen based on four basic taste conditions (sweet, bitter, sour, and salty). Participants rated their responses to each stimulus using both taste and emotion scales. The multivariate analyses including multidimensional scaling and classification analysis were performed. The findings revealed that auditory and gustatory stimuli in the sweet category were associated with positive valence, whereas those from the other taste categories were linked to negative valence. Additionally, auditory and gustatory stimuli in sour taste category were linked to high arousal, whereas stimuli in bitter taste category were associated with low arousal. This study revealed the potential mapping of gustatory and auditory stimuli onto core affect space in everyday experiences. Moreover, it demonstrated that emotions evoked by taste and music could be predicted across modalities, supporting modality-general representation of affect.

Anomaly detection in wind turbine blades based on PCA and convolutional kernel transform models: employing multivariate SCADA time series analysis

With the widespread application of wind power technology, the detection of abnormalities in wind turbine blades has become a key research area. The use of data from monitoring and data acquisition (SCADA) systems for data-driven fault detection research presents new challenges. This study utilizes short-term SCADA data from wind turbine generators to classify the blade abnormal and normal operational states, thereby introducing a new method called PCABSMMR. This strategy integrates principal component analysis (PCA) and borderline-synthetic minority over-sampling technique (Borderline-SMOTE) for data processing and utilizes an improved multi-dimensional time series classification (MTSC) model. It combines one-dimensional convolution from deep learning with shallow learning’s rigid classifiers. PCA is used for dimensionality reduction, while Borderline-SMOTE expands the samples of minority class fault instances. Comparative analysis with various methods shows that the proposed method has an average F1-score of 0.98, outperforming many state-of-the-art MTSC models across various evaluation metrics.

Level of Propensity: a zonal metric for evaluating the multimodal option Electric Micromobility and Public Transport

Electric Micromobility (EMM), e.g. e-bikes, e-scooters, light-weight electric vehicles, has a key role in policymakers’ strategies for making cities more liveable. Especially, its combination with Public Transport (PT) can be essential in the transportation sector decarbonization. This work proposes a zonal metric to evaluate EMM and PT integration, named “Level of Propensity” (LoP) to multimodality. LoP is defined adopting a multi-dimensional domain-based classification capturing two main phases of the trip: (1) the access phase by EMM towards PT stops; (2) the accessibility potential to destination by PT. PT open data, road network configuration, activities and Floating Car Data are adopted for the elaboration. LoP has been validated in the medium-size city of Salerno (Italy), but it is easily transferable to other cities. It can support urban planners and PT/EMM operators in identifying zones with low propensity to multimodality, suggesting priorities in terms of PT service improvements and EMM infrastructure implementations.

On the Classification of Regions of the Russian Federation by Level of Socio-Economic Development and Innovation Activity Index

The purpose of the study is to form an effective tool for classifying the regions of the Russian Federation in the context of innovation activity under technological constraints.Materials and methods. The information and empirical base of the research consists of Decrees of the President of the Russian Federation, normative legal acts of the Government of the Russian Federation, and open set of statistical data provided by the Federal State Statistics Service. The study provides a multidimensional classification of Russian regions using an indicative analysis of the innovative potential of territories and their socio-economic development, clustering by the Ward method and the Euclidean distance metric, as well as analysis, comparison and illustration of the results obtained using methods of visualizing information in tabular, graphical form, including the usage of cartograms.Results. While creating a model of typologization of Russian regions, a system of indicative indexes with both qualitative and quantitative characteristics was formed. Based on them, cluster analysis identifies five clusters of regions according to the level of innovation activity and socio-economic development, which is based on the final weighted average index (Iiv) of development. The results of the study determine the effectiveness of the author’s approach to the issue of typologization of regions, and emphasize the visibility and interactivity of the tools used within the framework of the model. The use of cluster and comparative analysis using weighted average indexes makes it possible to identify non-obvious patterns between the innovative activity of the region and its level of development, as well as to emphasize intuitively expected connections. All this can form a practical basis for developing effective strategies for regional development and improving the quality of life of the population.Conclusion. The study highlights the importance of analyzing the innovative activity of regions in modern conditions of sanctions wars and restrictions on technological imports. The developed model is a comprehensive tool for analyzing the innovative potential and level of development of regions, which allows identifying key factors of innovation activity and potential growth points, and thereby helping to form the basis for the development of strategies and programs for regional development aimed at improving the standard of living of the population.

Social Justice Profiles: An Exploratory Study towards an Empirically Based Multi-Dimensional Classification of Countries Regarding Fairness of Participation in Higher Education

The aim of this article is to suggest a better—theoretically and empirically grounded—understanding of the complex character of social justice in higher education. Theoretically, this article conceptualises social justice in higher education as mediating participation in, completion of and outcomes from higher education. It introduces the concept of composite capability for achieving higher education that captures capabilities to participate in, complete and gain outcomes from higher education. This study also develops a methodology for building an empirically based classification of countries regarding social justice in participation in higher education, taking into account the assessed inequality in students’ pathways to higher education as well as inequality in their social conditions, associated with students’ social origin. In so doing, it develops three indices: the index of inequalities in students’ pathways, the index of inequalities in students’ social conditions and the index of participation in higher education. Using microdata from the EUROSTUDENT VII survey (2019–2021) for 12 European countries, it applies the developed methodology to classify countries, for which data are available, by the degree of fairness in participation in higher education. This study’s results demonstrate the social embeddedness of social justice in higher education in different economic and political contexts.

Deep self-supervised machine learning algorithms with a novel feature elimination and selection approaches for blood test-based multi-dimensional health risks classification

BackgroundBlood test is extensively performed for screening, diagnoses and surveillance purposes. Although it is possible to automatically evaluate the raw blood test data with the advanced deep self-supervised machine learning approaches, it has not been profoundly investigated and implemented yet.ResultsThis paper proposes deep machine learning algorithms with multi-dimensional adaptive feature elimination, self-feature weighting and novel feature selection approaches. To classify the health risks based on the processed data with the deep layers, four machine learning algorithms having various properties from being utterly model free to gradient driven are modified.ConclusionsThe results show that the proposed deep machine learning algorithms can remove the unnecessary features, assign self-importance weights, selects their most informative ones and classify the health risks automatically from the worst-case low to worst-case high values.

Analytical Capabilities of Statistical Methodology for Assessing Threats to Economic Security

Analytical Capabilities of Statistical Methodology for Assessing Threats to Economic Security

Statistical analysis of the differentiation of regions of the Russian Federation by tax revenues of consolidated budgets using the R language

The subject area of this article is the application of descriptive statistics and multidimensional classification methods to describe the regional features of the formation of tax components of revenues of consolidated budgets of the Russian Federation. The aim of the work is to demonstrate the simplicity and effectiveness of using mathematical and statistical methods and the functionality of the open source language R to solve problems of structural analysis of tax revenues, identify regional specifics, and comparative analysis of regions from the point of view of tax revenues. The apparatus of mathematical statistics implemented in the R language, in particular, opens up wide opportunities for classifying tax subjects, including multidimensional ones, significantly facilitating the procedures of analysis, ranking and planning. The functionality described in the article can be used in the process of forming and adjusting tax policy at different levels. The possibilities of the mathematical statistics apparatus in combination with the instrumental methods of the R language are revealed by the example of the classification analysis of the regions of the Russian Federation. At the same time, the absolute and relative values of tax revenues in the revenues of regional budgets are selected as classification features. The classification by belonging to the federal district and the study of the "natural" stratification by the method of cluster analysis are considered. The apparatus of mathematical statistics and, especially, the tools of the R language are used unreasonably rarely in research of this kind, despite the ease of use and the absence of the need for special training, these circumstances determine the relevance of this article. Aggregation by federal districts made it possible to identify: the Ural Federal District as the leader in terms of the average regional share of tax revenues in the revenue part of the budget and the North Caucasus Federal District, characterized by the lowest average regional contributions of tax payments to regional budgets. The analysis of the natural stratification of the regions of the Russian Federation by their relative tax contributions to consolidated budgets made it possible to identify groups: the most typical regions, subsidized regions, donor regions and regions in which the most expensive assets of enterprises of the Russian Federation are concentrated

Partitioned 2D Set-Pruning Segment Trees with Compressed Buckets for Multi-Dimensional Packet Classification

Abstract Multi-dimensional packet classification is one of the most important functions to support various services in next generation routers. Both the memory-efficient data structure to support larger rule tables and the hardware architecture to achieve a higher throughput are desired. In this paper, we propose a parallel and pipelined architecture called Set-Pruning Segment Trees with Buckets (SPSTwB) for multi-dimensional packet classification. SPSTwB significantly reduces rule duplication based on a novel partitioning scheme and an efficient bucket merging scheme. The key feature of our proposed architecture is that memory consumption is reduced significantly regardless of the characteristics of various rule tables. In addition, the logic complexity of each pipeline stage is simplified by not storing rule IDs and priorities and thus it can run at a high clock rate. The proposed scheme needs less than 20 bytes per rule for various 100 K rule tables generated by ClassBench. In addition, the proposed scheme supports fast incremental rule update. The proposed pipelined architecture can achieve a throughput of 134 Gbps from the implementation on Xilinx Virtex-7 FPGA device with dual-ported Block RAM.

An Ensemble Deep Transfer Learning Model for Multi-dimensional Image Classification of Histological Prostate Biopsy Patterns

An Ensemble Deep Transfer Learning Model for Multi-dimensional Image Classification of Histological Prostate Biopsy Patterns

Testing the bipolar assumption of Singer-Loomis Type Deployment Inventory for Korean adults using classification and multidimensional scaling.

In this study, we explored whether the Korean version of Singer Loomis Type Deployment Inventory II (K-SLTDI) captures the opposing tendencies of Jung's theory of psychological type. The types are Extroverted Sensing, Extroverted Intuition, Extroverted Feeling, Extroverted Thinking, Introverted Sensing, Introverted Intuition, Introverted Feeling, and Introverted Thinking. A nationwide online survey was conducted in South Korea. We performed multidimensional scaling and classification analyses based on 521 Korean adult profiles with eight psychological types to test the bipolarity assumption. The results showed that the Procrustes-rotated four-dimensional space successfully represented four types of opposing tendencies. Moreover, the bipolarity assumption in the four dimensions of Jungian typology was tested and compared between lower and higher psychological distress populations via cluster analysis. Lastly, we explored patterns of responses in lower and higher psychological distress populations using intersubject correlation. Both similarity analyses and classification results consistently support the theoretical considerations on the conceptualization of Jung's type in independent order that the types could be derived without bipolar assumption as Singer and Loomis expected in their Type Development Inventory. Limitations in our study include the sample being randomly selected internet users during the COVID-19 pandemic, despite excellence in the use of the internet in the general Korean population.

Spatiotemporal changes of forest vegetation after the implementation of a natural forest protection project and underlying driving factors: Case study of a typical natural secondary forest area in the Loess Plateau

The Natural Forest Protection Project (NFPP) in China is the first large-scale ecological restoration project aimed at protecting natural forest resources in the world. However, since its implementation in 1998, there was a lack of accurate quantitative evaluation of the ecological governance and restoration effects. This study investigated the core forest area of Ziwuling on the Loess Plateau (LP). Using the Google Earth Engine (GEE) and the Landsat remote sensing dataset, a multidimensional classification feature set comprising spectrum + vegetation index + texture + terrain was constructed by extracting multitemporal remote sensing information. With the help of land use and vegetation type data, a sample database was established, after which the random forest algorithm was used to realize the automatic classification of the five periods of the study area in 2000, 2005, 2010, 2015 and 2020. Combined with geostatistical methods such as Theil-Sen (Sen) trend analysis, Mann-Kendall (MK) trend test and geographic detectors, we realized the trend analysis of the driving factors of vegetation growth and the contribution analysis of the spatial differentiation of different vegetation types. The results showed that: (1) From 2000 to 2020, the vegetation area increased by 141.92 km2. In terms of structural composition, the vegetation was dominated by hardwood forest, followed by sagebrush, and low coverage of coniferous forest. (2) Over time, the expansion of vegetation showed a positive slowing trend, and the growth rate in 2000–2005 was significantly higher than in 2005–2020. (3) The main driving force affecting vegetation restoration is precipitation, and the temperature plays a regulating role. (4) The main driving factors affecting the spatial differentiation of vegetation were annual maximum temperature, annual average temperature, and aspect. These findings not only help us understand the intuitive contribution of the NFPP to the vegetation restoration and ecological management of the LP, but also provide an important reference for future studies using remote sensing technology in ecological restoration monitoring.

The Method Radar: a way to organize methods for technology development with participation in mind

Abstract Identifying appropriate methods for any process, such as uncovering needs of juveniles in social welfare or designing interactive systems, requires intensive research and generally using a categorization system that brings methods in a systematic order. These taxonomy systems are heavily responsible for the later method usage and start the thinking process for researchers and practitioners alike in a specific direction. So making participation visible in such taxonomy systems directly affects the later method usage and makes participation more visible and easier to use. This article presents the Method Radar, a visualized categorization of methods with a focus on participation using the ladder of participation, that makes participation visible right at the beginning of any method selection. The Method Radar builds on the radar representation established in the technology sector, which allows a multi-dimensional classification. In addition, an implementation and systematic process for categorizing these methods are presented. It can be used for any form of method categorization in which participation is supposed to be thought of.

Supervised dimensionality reduction for multi-dimensional classification

Supervised dimensionality reduction for multi-dimensional classification

Diversity and Classification of Genetic Variations in Autism Spectrum Disorder

Autism spectrum disorder (ASD) is a neurodevelopmental condition with symptoms that affect the whole personality and all aspects of life. Although there is a high degree of heterogeneity in both its etiology and its characteristic behavioral patterns, the disorder is well-captured along the autistic triad. Currently, ASD status can be confirmed following an assessment of behavioral features, but there is a growing emphasis on conceptualizing autism as a spectrum, which allows for establishing a diagnosis based on the level of support need, free of discrete categories. Since ASD has a high genetic predominance, the number of genetic variations identified in the background of the condition is increasing exponentially as genetic testing methods are rapidly evolving. However, due to the huge amount of data to be analyzed, grouping the different DNA variations is still challenging. Therefore, in the present review, a multidimensional classification scheme was developed to accommodate most of the currently known genetic variants associated with autism. Genetic variations have been grouped according to six criteria (extent, time of onset, information content, frequency, number of genes involved, inheritance pattern), which are themselves not discrete categories, but form a coherent continuum in line with the autism spectrum approach.

Omics-CNN: A comprehensive pipeline for predictive analytics in quantitative omics using one-dimensional convolutional neural networks

Background and objectiveThe development of machine learning-based models that can be used for the prediction of severe diseases has been one of the main concerns of the scientific community. The current study seeks to expand a highly sophisticated tool, the Convolutional Neural Networks, making it applicable in multidimensional omics data classification problems and testing the newly introduced method on publicly available transcriptomics and proteomics data. MethodsIn this study, we introduce Omics-CNN, a Convolutional Neural Network-based pipeline, which couples Convolutional Neural Networks with dimensionality reduction, preprocessing, clustering, and explainability techniques to make them suitable to build highly accurate and interpretable classification models from high-throughput omics data. The developed tool has the potential to classify patients depending on the expression of genetic and clinical factors and identify features that can act as diagnostic biomarkers. Regarding dimensionality reduction, univariate and multivariate techniques were explored and compared. Gradient Weighted Class Activation Mapping analysis was performed to determine the most important features in the classification of the samples after training the model. ResultsThe newly introduced pipeline was applied to one transcriptomics and one proteomics dataset for the identification of diagnostic models and biosignatures for Ischemic Stroke (IS) and COVID-19 infection, reporting highly accurate biosignatures with accuracies of 96 % and 95.41 %, respectively. Meanwhile, classification models based solely on a small part of attributes provided lower predictive accuracy, but identified compact transcript biosignature (KRT15, VPRBP, TNFRSF4, GORASP2) for Ischemic Stroke and protein biosignature (ADGRB3, VNN2, AGER, CIAPIN1) for Covid-19 infection diagnosis, respectively. ConclusionsOmics-CNN, overcame the inherent problems of applying Convolutional Neural Networks for the training diagnostic models with quantitative omics data, outperforming previous models of machine learning developed using the same datasets for Ischemic Stroke and Covid-19 infection diagnosis, determining the most contributing biomarkers for both diseases.

Probability-based label enhancement for multi-dimensional classification

Multi-dimensional classification (MDC) assumes that each instance has multiple heterogeneous class spaces simultaneously, and each class variable describes the semantic information of instances from a specific dimension. Recent studies have proven that encoding heterogeneous class spaces into a special logical-label space and employing the label enhancement technique to learn latent real-number labels (i.e., label distributions) of instances is an effective strategy for MDC. However, the adopted label enhancement methods can result that data whose features are quite different to each other have similar label distributions. To tackle this problem, we propose a novel probability-based label enhancement approach for MDC. Specifically, manifold structures of the feature and label distribution spaces are transformed into two different probability distributions, and we expect them to be close. Subsequently, it makes label distributions of samples whose features have large differences be more differentiated. Moreover, the logical-label mapping and reconstruction terms are designed to preserve the intrinsic information from the logical-label space. Besides, an improved multi-output support vector regression is developed as the prediction model, where we introduce mean squared error to reduce the risk of model underfitting. Experimental results on ten benchmark datasets clearly validate the superiority of our method over state-of-the-art MDC baselines.

Receiver Operating Characteristic Optimization Based on Convex Hull and Evolutionary Algorithm

Classification is basically a Multi-objective problem. The efficiency of classification vastly depends on performance of a classifier which can be evaluated on the basis of Receiver Operating Characteristic (ROC) graph, Area under Curve (AUC), and selection of different threshold values are generally used as a tool. In machine learning, generally, 2-D classifiers are available that deal with bi-objective problems where overlapping of class may occur i.e. sensitivity and specificity may overlap. Recently, multi-class classification in which classes are mutually exclusive is in research trends along with the evolutionary algorithm. The application of ROC graph is extended to evaluate multi-dimensional classification as it is cost sensitive learning. The goal of this paper is to gather recent achievements in the field at one place to analyze the classifier performance for multi-dimensional classification problems using convex hull and evolutionary algorithms. In this paper, we tried to cover all the existing recent advance techniques in maximization of ROC and proposed a convex hull and evolutionary algorithm based new model for ROC maximization.

Assessing the impact of technological change on similar occupations: Implications for employment alternatives.

The fast-changing labor market highlights the need for an in-depth understanding of occupational mobility impacted by technological change. However, we lack a multidimensional classification scheme that considers similarities of occupations comprehensively, which prevents us from predicting employment trends and mobility across occupations. This study fills the gap by examining employment trends based on similarities between occupations. We first demonstrated a new method that clusters 756 occupation titles based on knowledge, skills, abilities, education, experience, training, activities, values, and interests. We used the Principal Component Analysis to categorize occupations in the Standard Occupational Classification, which is grouped into a four-level hierarchy. Then, we paired the occupation clusters with the occupational employment projections provided by the U.S. Bureau of Labor Statistics. We analyzed how employment would change and what factors affect the employment changes within occupation groups. Particularly, we specified factors related to technological changes. The results reveal that technological change accounts for significant job losses in some clusters. This poses occupational mobility challenges for workers in these jobs at present. Job losses for nearly 60% of current employment will occur in low-skill, low-wage occupational groups. Meanwhile, many mid-skilled and highly skilled jobs are projected to grow in the next ten years. Our results demonstrate the utility of our occupational classification scheme. Furthermore, it suggests a critical need for skills upgrading and workforce development for workers in declining jobs. Special attention should be paid to vulnerable workers, such as older individuals and minorities.

Classification of Russian Regions by the Level of Agricultural Development in 2019–2021

The urgency of improving mathematical and statistical tools for analyzing agricultural complex across the territories of our country depends on the gravity of the issue of interregional differentiation of the scale of development of agricultural production, which plays a crucial role in the implementation of the Food Security Doctrine of the Russian Federation.The introductory part of the article presents an overview of the literature on food security issues, an analysis of the current state of domestic agriculture, and directions for its development. The body of the article explains the research methodology, which allows to classify, first andforemost, agricultural regions (indicating their territorial location, socio-economic characteristics, and predictive characteristics). As an indicator of the intensity of agricultural development in the regions, the authors used the volume of agricultural production per capita, which became a criterion for selecting 25% of the regions with the maximum values of this indicator. That said, by an agro-region (or agricultural region) is meant a region that produces more agricultural production than needed and exports its surplus outside the region or country (food exports). In the multidimensional classification of agro-regions, three clusters are determined (according to the level of socio-economic development, the development efficiency of the economic activity «Agriculture, forestry, hunting, and fishing», and the volume of crop and livestock production). The paper analyses the dynamics and predictive characteristics of the development of the cluster-average volume of agricultural production per capita (including crop production and animal husbandry). An analysis of the dynamics of exports of food products and agricultural raw materials per capita by clusters revealed that the regions of the first cluster are the most actively developing, and the regions of the third cluster (with the most developed animal husbandry) are oriented towards domestic consumers.According to the authors, during the analyzed period in Russia, there was a positive trend in the development of agriculture, with the development of agro-regions outpacing all-Russian trends. The highest development rates are typical for agro-regions with a balanced development of crop and livestock production (regions of the first cluster).