A description is provided of a method -- the Set Equality Method (SEM) -- to determine the validity, or lack of validity, of each categorical syllogism. A justification is given for the presentation of a new method to solve a problem which has already been solved using different approaches. First, the SEM assigns an equality of certain sets (or two of those equalities in specific cases as will be indicated) to each of the categorical propositions composing each syllogism considered -- that is, to each of the two premises and to the conclusion. Each syllogism considered is valid if and only if a) it is possible to select one of those equalities corresponding to one of the premises such that one of its members is a certain set and the other of those equalities corresponding to the other premise such that one of its members is a subset of the set mentioned, and b) it is possible to deduce an equality corresponding to the conclusion of the two equalities corresponding to the premises. In some cases, as will be specified, it is possible to provide a second test for the validity of a syllogism whose validity was already proven, thus providing information about the logical form of categorical syllogisms.

An information-theoretic data mining method is employed to analyze categorical spatiotemporal Geographic Information System land use data. Reconstructability Analysis (RA) is a maximum-entropy-based data modeling methodology that works exclusively with discrete data such as those in the National Land Cover Database (NLCD). The NLCD are organized into a spatial (raster) grid, and are available in consistent format for every five years from 2001 to 2021. An NLCD tool reports how much change occurred for each category of land use; for the study area examined the most dynamic class is Evergreen Forest (EFO), so the presence or absence of EFO in 2021 was chosen as the dependent variable that our data modeling attempts to predict. RA predicts the outcome with approximately 80% accuracy using a sparse set of cells from a spacetime data cube consisting of neighboring lagged-time cells. When the predicting cells are all Shrubs & Grasses there is a high probability for a 2021 state of EFO, while when the predicting cells are all EFO, there is a high probability that the 2021 state will be not-EFO. These findings are interpreted as detecting forest clear-cut cycles that show up in the data, and explain why this class is so dynamic. This study introduces a new approach to analyzing GIS categorical data and expands the range of applications that this entropy-based methodology can successfully model.

Machine Learning (ML) methods have become important to enhance the performance of decision-support predictive models. However, class imbalance is one of the main challenges for developing ML models, because it limits the generalization of these models, and biases the learning algorithms. In this paper, we consider oversampling methods for generating synthetic categorical clinical data aiming to improve the predictive performance in ML models, and the identification of risk factors for cardiovascular diseases (CVDs). We performed a comparative study of several categorical synthetic data generation methods, including Generative Adversarial Networks (GANs). Then, we assessed the impact of combining oversampling strategies and linear and nonlinear supervised ML methods. Lastly, we conducted a post-hoc model interpretability based on the importance of the risk factors. Experimental results show the potential of GAN-based models for generating high-quality categorical synthetic data, yielding probability mass functions that are highly close to real data, maintaining relevant insights, and contributing to increase the predictive performance. The GAN-based model and a linear classifier outperforms other oversampling techniques, improving 2\% the area under the curve. These results demonstrate the capability of synthetic data to help both in determining risk factors and building models for CVD prediction.

We prove that a diffeological space is diffeomorphic to the categorical join of any generating family of plots.

Conditional Independence (CI) testing is fundamental in statistical analysis. For example, CI testing helps validate causal graphs or longitudinal data analysis with repeated measures in causal inference. CI testing is difficult, especially when testing involves categorical variables conditioned on a mixture of continuous and categorical variables. Current parametric and non-parametric testing methods designed for continuous variables and can quickly fall short in the categorical case. This paper presents a computational approach for CI testing for categorical data types, which we call computational conditional independence (CCI) testing. The test procedure is based on permutation and combines machine learning prediction algorithms and Monte Carlo Cross-Validation. We evaluated the approach through simulation studies and assessed the performance against alternative methods; the generalized covariance measure (GCM) test, the kernel conditional independence (KCI) test, and testing with multinomial regression. We find that the computational approach to testing has utility over the alternative methods, and we can achieve better control over type-I-error rates. We hope this work can expand the toolkit for CI testing for practitioners and researchers.

The Republic of Djibouti is a small country in the Horn of Africa and, as in most developing countries, rain gauges are sparse and data are scarce. This study aims to report on the reliability of gridded precipitation datasets (P datasets) across the Republic of Djibouti through direct comparisons with rain gauge measurements from the annual to the daily time scales. Our specific objective is to be able to use such products in the context of hydrological modeling at a daily time step. Given the scarcity of available data in the Republic of Djibouti, our study was carried out on two time windows (1980-1990 and 2008-2013) and two gauge networks with different spatial resolutions: the southeast of the Republic of Djibouti (5000 km2) and the Ambouli catchment (794 km2), which drains the city of Djibouti. The reliability of these products is analyzed with quantitative metrics and categorical metrics, exclusively at a daily time step for the latter. The performance of the P datasets degrades from the annual time scale to the daily time scale. Even though the same products exhibit the best performance at the various time scales, the performance of most of the products differs from one spatial scale to another. Our results demonstrate the importance of the temporal and spatial windows, as the same products can perform differently according to the scale. For all the spatiotemporal scales, the most reliable product is MSWEP v.2.2. This P dataset is derived from a combination of satellite products (multiple sensors such as infrared and passive microwave), reanalysis products, and rain gauge observations. A strong discrepancy between rain gauge observations and P datasets is revealed according to the categorical metric at a daily time step. The analysis of rainfall events triggering runoff, using a 10 mm rainfall threshold showed that the most efficient products were unable to accurately detect such events at a daily time step, with a significant underestimation of rainfall events higher than 10 mm. None of these products, even the most reliable, can be used for a calibration/validation of a hydrological model at a daily time step.

This paper offers a comprehensive examination of the process involved in developing and automating supervised end-to-end machine learning workflows for forecasting and classification purposes. It offers a complete overview of the components (i.e. feature engineering, model selection, etc), principles (i.e. bias-variance decomposition, model complexity, overfitting, model sensitivity to feature assumptions and scaling, output interpretability, etc), models (i.e. neural networks, regression models, etc), methods (i.e. Cross-Validation, data augmentation, etc), metrics (i.e. Mean Squared Error, F1-score, etc) and tools that rule most supervised learning applications with numerical and categorical data, as well as their integration, automation, and deployment. The end goal and contribution of this paper is the education and guidance of the non-AI expert academic community over complete and rigorous machine learning pipelines and data science practices, from problem scoping to design and state-of-the-art automation tools, including basic principles and reasoning in the choice of methods. The paper delves into the critical stages of supervised machine learning workflow development, many of which are often omitted by researchers due to brevity, and covers foundational concepts essential for understanding and optimizing a functional machine learning workflow, thereby offering a holistic view of task-specific application development for applied researchers who are not AI experts.

Speech emotion recognition is a critical component for achieving natural human-robot interaction. The modulation-filtered cochleagram is a feature based on auditory modulation perception, which contains multi-dimensional spectral-temporal modulation representation. In this study, we propose an emotion recognition framework that utilizes a multi-level attention network to recognize emotions from the modulation-filtered cochleagram. The channel-level attention and spatial-level attention modules are used to capture emotional saliency maps of channel and spatial feature representations from the 3D convolution feature maps, respectively. Furthermore, the temporal-level attention module captures significant emotional regions from the concatenated feature sequence of the emotional saliency maps. Our experiments on the IEMOCAP dataset demonstrate that the modulation-filtered cochleagram significantly improves the prediction performance of categorical emotion compared to other evaluated features. Moreover, our emotion recognition framework achieves a better unweighted accuracy of 71% in categorical emotion recognition than several existing approaches in the experiments. In summary, our study demonstrates the effectiveness of the modulation-filtered cochleagram in speech emotion recognition, and our proposed multi-level attention framework provides a promising direction for future research in this field.

A review of literature, supporting the view on the psychophysical origins of some acceptance effects of cyber-physical systems (CPSs), is presented and discussed in the paper. Psychological effects like the reactance to a robot or the uncanny valley phenomenon suggest that CPSs are perceived as a special type of ‘being’, a phylogenetically emergent category, or a novel ontology in a philosophical sense, culturally developing as a function of the social brain. Justification of this view is provided by the psychophysically-relevant human responses to technologically and socially ambiguous stimuli on the one hand, and the probabilistic independence of the proposed ontology to similar ontologies, subjectively experienced by the human, on the other. The results of the presented in the paper analysis demonstrate the feasibility of the proposed hypothesis of ontological ‘near independence’ of the distribution of the probabilistic psychophysical processes in response to the cyber-physical stimulus as a socially evolving phenomenon.

Among the various generalizations of persistent topology, the one based on rank functions and leading to indexing-aware functions appears to be particularly suited to catch graph-theoretical properties without the need for a simplicial construction and a homology computation. This paper defines and studies "simple" and "single-vertex" features in directed and undirected graphs, by which several indexing-aware persistence functions are produced, within the scheme of steady and ranging sets. The implementation of the "sink" feature and its application to trust networks provide an example of the ease of use and meaningfulness of the method.

Objectives: Although pitch reception poses a great challenge for individuals with cochlear implants (CIs), formal auditory training (e.g., high variability phonetic training, HVPT) has been shown to provide direct benefits in pitch-related perceptual performances such as lexical tone recognition for CI users. As lexical tones in spoken language are expressed with a multitude of distinct spectral, temporal, and intensity cues, it is important to determine the sources of training benefits for CI users. The purpose of the present study was to conduct a rigorous fine-scale evaluation with the categorical perception (CP) paradigm to control the acoustic parameters and test the efficacy and sustainability of HVPT for Mandarin-speaking pediatric CI recipients. The main hypothesis was that HVPT-induced perceptual learning would greatly enhance CI users’ ability to extract the primary pitch contours from spoken words for lexical tone identification and discrimination. Furthermore, individual differences in immediate and long-term gains from training would likely be attributable to baseline performance and duration of CI use. Design: Twenty-eight prelingually deaf Mandarin-speaking kindergarteners with CIs were tested. Half of them received five sessions of HVPT within a period of three weeks. The other half served as control who did not receive the formal training. Two classical CP tasks on a tonal continuum from Mandarin Tone 1 (high-flat in pitch) to Tone 2 (mid-rising in pitch) with fixed acoustic features of duration and intensity were administered before (pretest), immediately after (posttest), and 10 weeks post training termination (follow-up test). Participants were instructed to either label a speech stimulus along the continuum (i.e., identification task) or determine whether a pair of stimuli separated by zero or two steps from the continuum was the same or different (i.e., discrimination task). Identification function measures (i.e., boundary position and boundary width) and discrimination function scores (i.e., between-category score, within-category score, and peakedness score) were assessed for each child participant across the three test sessions.Results: Linear mixed-effects (LME) models showed significant training-induced enhancement in lexical tone categorization with significantly narrower boundary width and better between-category discrimination in the immediate posttest over pretest for the trainees. Furthermore, training-induced gains were reliably retained in the follow-up test 10 weeks after training. By contrast, no significant changes were found in the control group across sessions. Regression analysis confirmed that baseline performance (i.e., boundary width in the pretest session) and duration of CI use were significant predictors for the magnitude of training-induced benefits. Conclusions: The stringent CP tests with synthesized stimuli that excluded acoustic cues other than the pitch contour and were never used in training showed strong evidence for the efficacy of HVPT in yielding immediate and sustained improvement in lexical tone categorization for Mandarin-speaking children with CIs. The training results and individual differences have remarkable implications for developing personalized computer-based short-term HVPT protocols that may have sustainable long-term benefits for aural rehabilitation in this clinical population.

It is very difficult to accurately divide farmland and woodland in Gaofen 2 (GF-2) remote sensing image, because their single plant coverage is very small, and their spectra are very similar. The ratio of spatial resolution and one plant’s coverage area must be fully taken into account when designing the Convolutional Neural Network structure for extracting them from GF-2 image. We establish a Convolutional Encode Neural Networks model (CENN), The first layer has two sets of convolution kernels to learn the characteristics of farmland and woodland respectively, while the second layer is the encoder to encode the characteristics by transfer function, which can map the results to the corresponding category number. In the training stage, samples of farmland, woodland, and other categories are categorically used to train CENN, as soon as training is accomplished, CENN would acquire enough ability to accurately extract farmland and woodland from remote sensing images. The final extraction result is obtained by implementing per-pixel segmentation of images used to train the CENN. CENN is compared and analyzed with others such as Deep Belief Network (DBN), Full Convolutional Network (FCN), Deeplab Model. The results of experiments show that CENN can more accurately mine the characteristics of farmland and woodland, and it achieves its goal of extracting farmland and woodland with high precision from GF-2 images.