All issues

text-align: justify;">In this paper the problem of automated building of a citation graph from a collection of scientific documents is considered as a sequence of machine learning tasks. The overall data processing technology is described which consists of six stages: preprocessing, metainformation extraction, bibliography lists extraction, splitting bibliography lists into separate bibliography records, standardization of each bibliography record, and record linkage. The goal of this paper is to provide a survey of approaches and algorithms suitable for each stage, motivate the choice of the best combination of algorithms, and adapt some of them for multilingual bibliographies processing. For some of the tasks new algorithms and heuristics are proposed and evaluated on the mixed English and Russian documents corpora.

text-align: justify;">The paper studies the problem of searching for fragments with missing bibliographic links in a scientific article using automatic binary classification. To train the model, we propose a new contrast resampling technique, the innovation of which is the consideration of the context of the link, taking into account the boundaries of the fragment, which mostly affects the probability of presence of a bibliographic links in it. The training set was formed of automatically labeled samples that are fragments of three sentences with class labels «without link» and «with link» that satisfy the requirement of contrast: samples of different classes are distanced in the source text. The feature space was built automatically based on the term occurrence statistics and was expanded by constructing additional features — entities (names, numbers, quotes and abbreviations) recognized in the text.

text-align: justify;">A series of experiments was carried out on the archives of the scientific journals «Law enforcement review» (273 articles) and «Journal Infectology» (684 articles). The classification was carried out by the models Nearest Neighbors, RBF SVM, Random Forest, Multilayer Perceptron, with the selection of optimal hyperparameters for each classifier.

text-align: justify;">Experiments have confirmed the hypothesis put forward. The highest accuracy was reached by the neural network classifier (95%), which is however not as fast as the linear one that showed also high accuracy with contrast resampling (91–94%). These values are superior to those reported for NER and Sentiment Analysis on comparable data. The high computational efficiency of the proposed method makes it possible to integrate it into applied systems and to process documents online.

text-align: justify;">The technology of creating patterns from natural language words (concepts) based on text data in the bag of words model is considered. Patterns are used to reduce the dimension of the original space in the description of documents and search for semantically related words by topic. The process of dimensionality reduction is implemented through the formation of patterns of latent features. The variety of structures of document relations is investigated in order to divide them into themes in the latent space.

text-align: justify;">It is considered that a given set of documents (objects) is divided into two non-overlapping classes, for the analysis of which it is necessary to use a common dictionary. The belonging of words to a common vocabulary is initially unknown. Class objects are considered as opposition to each other. Quantitative parameters of oppositionality are determined through the values of the stability of each feature and generalized assessments of objects according to non-overlapping sets of features.

text-align: justify;">To calculate the stability, the feature values are divided into non-intersecting intervals, the optimal boundaries of which are determined by a special criterion. The maximum stability is achieved under the condition that the boundaries of each interval contain values of one of the two classes.

text-align: justify;">The composition of features in sets (patterns of words) is formed from a sequence ordered by stability values. The process of formation of patterns and latent features based on them is implemented according to the rules of hierarchical agglomerative grouping.

text-align: justify;">A set of latent features is used for cluster analysis of documents using metric grouping algorithms. The analysis applies the coefficient of content authenticity based on the data on the belonging of documents to classes. The coefficient is a numerical characteristic of the dominance of class representatives in groups.

text-align: justify;">To divide documents into topics, it is proposed to use the union of groups in relation to their centers. As patterns for each topic, a sequence of words ordered by frequency of occurrence from a common dictionary is considered.

text-align: justify;">The results of a computational experiment on collections of abstracts of scientific dissertations are presented. Sequences of words from the general dictionary on 4 topics are formed.

text-align: justify;">The method of the stochastic matrix spectrum analysis is used to build an indicator that allows to determine the subject of scientific texts without keywords usage. This matrix is a matrix of conditional probabilities of bigrams, built on the statistics of the alphabet characters in the text without spaces, numbers and punctuation marks. Scientific texts are classified according to the mutual arrangement of invariant subspaces of the matrix of conditional probabilities of pairs of letter combinations. The separation indicator is the value of the cosine of the angle between the right and left eigenvectors corresponding to the maximum and minimum eigenvalues. The computational algorithm uses a special representation of the dichotomy parameter, which is the integral of the square norm of the resolvent of the stochastic matrix of bigrams along the circumference of a given radius in the complex plane. The tendency of the integral to infinity testifies to the approximation of the integration circuit to the eigenvalue of the matrix. The paper presents the typical distribution of the indicator of identification of specialties. For statistical analysis were analyzed dissertations on the main 19 specialties without taking into account the classification within the specialty, 20 texts for the specialty. It was found that the empirical distributions of the cosine of the angle for the mathematical and Humanities specialties do not have a common domain, so they can be formally divided by the value of this indicator without errors. Although the body of texts was not particularly large, nevertheless, in the case of arbitrary selection of dissertations, the identification error at the level of 2 % seems to be a very good result compared to the methods based on semantic analysis. It was also found that it is possible to make a text pattern for each of the specialties in the form of a reference matrix of bigrams, in the vicinity of which in the norm of summable functions it is possible to accurately identify the theme of the written scientific work, without using keywords. The proposed method can be used as a comparative indicator of greater or lesser severity of the scientific text or as an indicator of compliance of the text to a certain scientific level.

text-align: justify;">In this paper the various methods of machine classification of scientific texts by thematic sections on the example of publications in specialized medical journals published by Springer are compared. The corpus of texts was studied in five sections: pharmacology/toxicology, cardiology, immunology, neurology and oncology. We considered both classification methods based on the analysis of annotations and keywords, and classification methods based on the processing of actual texts. Methods of Bayesian classification, reference vectors, and reference letter combinations were applied. It is shown that the method of classification with the best accuracy is based on creating a library of standards of letter trigrams that correspond to texts of a certain subject. It is turned out that for this corpus the Bayesian method gives an error of about 20%, the support vector machine has error of order 10%, and the proximity of the distribution of three-letter text to the standard theme gives an error of about 5%, which allows to rank these methods to the use of artificial intelligence in the task of text classification by industry specialties. It is important that the support vector method provides the same accuracy when analyzing annotations as when analyzing full texts, which is important for reducing the number of operations for large text corpus.