Using coexpression to explore cell-type diversity with the fcoex package

Here, we present the fcoex package, which infers coexpression from scRNA-seq data and yields multiple, overlapping classes of cells based on coexpression modules. The tool extends the current scRNA-seq toolbox, providing a multi-hierarchy view on cell functionality and enabling the development of more complete cell atlases. Single-cell RNA sequencing (scRNA-seq) captures details of the cellular landscape, basing a fine-grained view on biological processes. Current pipelines, however, are restricted to single-label perspectives, missing details of the classification landscape. In the pbmc3k blood cell dataset, fcoex detects known classes, like antigen-presenting cells and a new theoretical group of cells, marked by the expression of FCGR3A (CD16). Fcoex extends the current scRNA-seq toolbox, providing a multi-hierarchy view on cell functions as a tool to develop complete cell type atlases.

Adaptive Probabilistic Neuro-Fuzzy System and its Hybrid Learning in Medical Diagnostics Task

Background: The medical diagnostic task in conditions of the limited dataset and overlapping classes is considered. Such limitations happen quite often in real-world tasks. The lack of long training datasets during solving real tasks in the problem of medical diagnostics causes not being able to use the mathematical apparatus of deep learning. Additionally, considering other factors, such as in a dataset, classes can be overlapped in the feature space; also data can be specified in various scales: in the numerical interval, numerical ratios, ordinal (rank), nominal and binary, which does not allow the use of known neural networks. In order to overcome arising restrictions and problems, a hybrid neuro-fuzzy system based on a probabilistic neural network and adaptive neuro-fuzzy interference system that allows solving the task in these situations is proposed. Methods: Computational intelligence, artificial neural networks, neuro-fuzzy systems compared to conventional artificial neural networks, the proposed system requires significantly less training time, and in comparison with neuro-fuzzy systems, it contains significantly fewer membership functions in the fuzzification layer. The hybrid learning algorithm for the system under consideration based on self-learning according to the principle “Winner takes all” and lazy learning according to the principle “Neurons at data points” has been introduced. Results: The proposed system solves the problem of classification in conditions of overlapping classes with the calculation of the membership levels of the formed diagnosis to various possible classes. Conclusion: The proposed system is quite simple in its numerical implementation, characterized by a high speed of information processing, both in the learning process and in the decision-making process; it easily adapts to situations when the number of diagnostics features changes during the system's functioning.

Capability Language Processing (CLP): Classification and Ranking of Manufacturing Suppliers Based on Unstructured Capability Data

The unstructured data available on the websites of manufacturing suppliers can provide useful insights into the technological and organizational capabilities of manufacturers. However, since the data is often represented in an unstructured form using natural language text, it is difficult to efficiently search and analyze the capability data and learn from it. The objective of this work is to propose a set of text analytics techniques to enable automated classification and ranking of suppliers based on their capability narratives. The supervised classification and semantic similarity measurement methods used in this research are supported by a formal thesaurus that uses SKOS (Simple Knowledge Organization System) for its syntax and semantics. Normalized Google Distance (NGD) was used as a metric for measuring the relatedness of terms. The proposed framework was validated experimentally using a hypothetical search scenario. The results indicate that the generated ranked list shows a high correlation with human judgment specially if the query concept vector and supplier concept vector belong to the same class. However, the correlation decreases when multiple overlapping classes of suppliers are mixed together. The findings of this research can be used to improve the precision and reliability of Capability Language Processing (CLP) tools and methods.

Input Output for Document Classifier

The report generated displays a list of automatically generated keywords in each document. A document is allowed to have any number of keywords. As the keywords are getting generated at any pass of the loop, there is no restriction on the width of keywords. Another report is also generated to display the list of the document class. If a document finds its match with more than one class (overlapping classes), the selection of the final class for a document is done on the basis of the maximum weight of the keywords in each class.

Stanislav Voronin’s Universal Classification of Onomatopoeic Words: a Critical Approach (Part 1)

Introduction. The universal classification of onomatopoeic words was first introduced in 1969 by Stanislav V. Voronin. In the course of the following fifty years it has been tested on the material of typologically different languages both by the author himself and by other researchers. The aim of this article is to provide a full description of the classification (which has never been published in English before) and to examine its key points critically. The bulk of empirical data collected in the recent years calls for yet another update on the classification. There is a logical contradiction between such classes of onomatopoeic words as frequentatives and frequentatives-(quasi)-instants-continuants. They overlap typologically. This and other minor issues are solved in the present paper.Methodology and sources. The method discussed and applied in the classification is the method of phonosemantic analysis introduced by S. V. Voronin. Empirical data from English and other relevant languages are used for supporting the proposed changes into the classification.Results and discussion. The critical analysis of the Voronin’s universal classification of the onomatopoeic words revealed the presence of overlapping classes and hyperclasses within it, as well as other minor inconsistencies. The empirical typological data allowed to introduce some minor corrections while retaining the main principles of the classification.Conclusion. Introduced half a century ago, Stanislav Voronin’s classification of onomatopoeic words still remains a useful tool of typological research. Critical additions and proposed changes do not lessen its impact on studies in linguistic iconicity. The first part of this paper is devoted to the description of the classification and to the discussion of its advantages and limitations. In the second part of the article some possible solutions to the detected problems are suggested.

The ‘nouniness’ of attributive adjectives and ‘verbiness’ of predicative adjectives: evidence from phonology

This article investigates prototypically attributive versus predicative adjectives in English in terms of the phonological properties that have been associated especially with nouns versus verbs in a substantial body of psycholinguistic research (e.g. Kelly 1992) – often ignored in theoretical linguistic work on word classes. Inspired by Berg's (2000, 2009) ‘cross-level harmony constraint’, the hypothesis I test is that prototypically attributive adjectives not only align more with nouns than with verbs syntactically, semantically and pragmatically, but also phonologically – and likewise for prototypically predicative adjectives and verbs. I analyse the phonological structure of frequent adjectives from the Corpus of Contemporary American English (COCA), and show that the data do indeed support the hypothesis. Berg's ‘cross-level harmony constraint’ may thus apply not only to the entire word classes noun, verb and adjective, but also to these two adjectival subclasses. I discuss several theoretical issues that emerge. The facts are most readily accommodated in a usage-based model, such as Radical Construction Grammar (Croft 2001), where these adjectives are seen as forming two distinct but overlapping classes. Drawing also on recent research by Boyd & Goldberg (2011) and Hao (2015), I explore the possible nature and emergence of these classes in some detail.