Self organizing maps (SOMs) or cognitive maps are designed to preserve major topological attributes of manifolds in a higher dimensionality data space to corresponding projections thereof in a low dimensional coordinate space. This is performed by mapping neighborhoods and the distances contained therein from the data space to ones in the coordinate space. Thus, SOM functionality relies heavily on the geometrical properties of both spaces. Topologically flexible data space distance metrics are constructed by combining tensors with graph filters, the latter coming from graph signal processing. The power of these distance metrics comes from naturally expressing the higher order relationships between points of the data space. This paves the way for addressing engineering scenarios involving a large number of densely interrelated attributes. One such case is discerning personalities from textual information based on the Myers-Briggs taxonomy indicator (MBTI), a framework of archetypal personalities derived from Jungian psychodynamic theory. Various graph filters were tested on a benchmark Kaggle dataset with ground truth with comparisons assessed in terms of topological error, cluster purity, average inter-cluster distance, and cluster curvature variability. Data points were stored in Pinecone, a recent vector database, with Python integration.

In today’s digital era, the internet is an indispensable platform for self-expression, facilitating communication, idea sharing, and community formation. Language, a pivotal tool in these online interactive spaces, is vital in reflecting personal identities, notably gender identification. This paper investigates gender identification on online discussion platforms, recognizing the crucial role of language in reflecting personal identities. The study employs Natural Language Processing techniques and machine learning algorithms to analyze data from a public discussion website. Beginning with a comprehensive literature review, the research explores the nexus between gender and language in online and offline contexts. The methodology involves data gathering, extensive preprocessing, and in-depth exploratory analysis, employing statistical methods and graphical representations. The study then rigorously evaluates their accuracy and effectiveness by applying diverse algorithms and models for gender-based text categorization. Results indicate the superior performance of transformer models, particularly distilBERT, in categorizing gender accurately. Additionally, the research underscores the challenges of gender-neutral analysis, emphasizing the need for inclusive methodologies in non-binary gender classification. The study contributes to the broader field of gender studies, providing valuable insights for future research and discussions on the interplay of gender and language in online spaces.

Matrix factorization is paramount in large scale graph mining as well as a versatile paradigm for dimensionality reduction. In particular, factoring a graph adjacency matrix may well reveal, depending on the specific f actor properties, higher order structure. The latter describes global graph properties better compared to first order connectivity patterns such as vertex degrees. The Boolean semiring factorization of an adjacency matrix yields a product of two smaller and sparser matrices where the former contains disjoint fundamental vertex subsets and the latter combinations thereof. Therefore, the first factor represents community structure and the second has the cross connections between them. In this way graph partitioning and dimensionality reduction are simultaneously achieved. Because of the nature of the Boolean semiring, most common linear algebraic solvers cannot be applied. Moreover, the exact factorization is NP hard. To address these limitations, a genetic algorithm has been developed with evolutionary operations tailored to heuristically compute said factorization which offers interpretability and a high parallelism potential. Besides graph mining the major applications of the Boolean semiring factorization include role mining in enterprise database and operating system realms, curriculum design, and graph flows under inflexible uniqueness constraints. The results obtained by applying the proposed genetic algorithm to synthetic graph benchmarks are very encouraging.

Social networks are an essential component of people’ daily lives, and as a result, much academic attention has been focused on them. The rapid adoption of machine learning as a problem-solving tool, which simplifies and accelerates numerous tasks while enabling the processing of large volumes of data, has played a significant role in this field of research. This is in contrast to the more traditional approaches that lacked this momentum. Characterization of linkages and cluster identification i n social networks are two of the research community’s most well-known issues. The goal of this study is to gather data for a set of users who are then divided into groups based on the hashtags they used in their Twitter postings. The procedure performed generates the numerical data, in following reduces the dimensions, and finally performs the clustering.

At the dawn of the Internet era graph analytics play an important role in high- and low-level network policymaking across a wide array of fields so diverse as transportation network design, supply chain engineering and logistics, social media analysis, and computer communication networks, to name just a few. This can be attributed not only to the size of the original graph but also to the nature of the problem parameters. For instance, algorithmic solutions depend heavily on the approximation criterion selection. Moreover, iterative or heuristic solutions are often sought as it is a high dimensional problem given the high number of vertices and edges involved as well as their complex interaction. Replacing under constraints a directed graph with an undirected one having the same vertex set is often sought in applications such as data visualization, community structure discovery, and connection-based vertex centrality metrics. Polar decomposition is a key matrix factorization which represents a matrix as a product of a symmetric positive (semi)definite factor and an orthogonal one. The former can be an undirected approximation of the original adjacency matrix. The proposed graph approximation has been tested with three Twitter graphs with encouraging results with respect to density, Fiedler number, and certain vertex centrality metrics based on matrix power series. The dataset was hosted in an online MongoDB instance.

Graph neural networks (GNNs) is an emerging class of iterative connectionist models taking full advantage of the interaction patterns in an underlying domain. Depending on their configuration GNNs aggregate local state information to obtain robust estimates of global properties. Since graphs inherently represent high dimensional data, GNNs can effectively perform dimensionality reduction for certain aggregator selections. One such task is assigning sentiment polarity labels to the vertices of a large social network based on local ground truth state vectors containing structural, functional, and affective attributes. Emotions have been long identified as key factors in the overall social network resiliency and determining such labels robustly would be a major indicator of it. As a concrete example, the proposed methodology has been applied to two benchmark graphs obtained from political Twitter with topic sampling regarding the Greek 1821 Independence Revolution and the US 2020 Presidential Elections. Based on the results recommendations for researchers and practitioners are offered.

In this article, an approach to semantic image analysis is presented. Under the proposed approach, ontologies are used to capture general, spatial, and contextual knowledge of a domain, and a genetic algorithm is applied to realize the final annotation. The employed domain knowledge considers high-level information in terms of the concepts of interest of the examined domain, contextual information in the form of fuzzy ontological relations, as well as low-level information in terms of prototypical low-level visual descriptors. To account for the inherent ambiguity in visual information, uncertainty has been introduced in the spatial relations definition. First, an initial hypothesis set of graded annotations is produced for each image region, and then context is exploited to update appropriately the estimated degrees of confidence. Finally, a genetic algorithm is applied to decide the most plausible annotation by utilizing the visual and the spatial concepts definitions included in the domain ontology. Experiments with a collection of photographs belonging to two different domains demonstrate the performance of the proposed approach.