An AI-Based Methodology for the Automatic Classification of a Multiclass Ebook Collection Using Information From the Tables of Contents

Abstract

Book recommendation to support professors and students in the identification of relevant sources is of significant importance for both universities and digital libraries and, hence, motivates the development of a recommendation system. This paper aims at automatically classifying a multiclass corpus that was created from ebooks from the Springer collection, which is available through the Hellenic Academic Libraries' subscription, by utilizing an unsupervised neural network (NN) (self-organizing maps, SOM) and two deep neural network (DNN) architectures, namely, a long short-term memory (LSTM) and a convolutional neural network (CNN) combined with a LSTM(CNN+LSTM) under various configuration scenarios. The vector construction leverages information that was extracted from the table of contents (ToC) of each book using the TF-IDF weighting scheme (for the first case) and the Keras tokenizer (for the second). Extensive experiments were conducted using various configurations of preprocessing steps, NN set up and vector and vocabulary sizes to assess their impact on the classifier's performance. Furthermore, we show that majority voting is more suitable for selecting the dominant label for a specified node. The experimental analysis showed the feasibility of developing a recommendation system for supporting professors and students in the identification of related sources based on a detailed thematic description (e.g., abstract or table of contents of a book) rather than a few keywords. In the conducted experiments, the subsystem that utilized the DNN (LSTM) performed the best, with F1-scores of 67% for the 26 categories and 80% for the 5 general categories, whereas SOM realizes F1-scores of less than 5% in both cases.

Multisubject Analysis and Classification of Books and Book Collections, Based on a Subject Term Vocabulary and the Latent Dirichlet Allocation

“…In [11] two different AI approaches to assigning labels to documents are reported, namely a Deep Neural Network (DNN) and a Self-Organizing Maps (SOM). The DNN uses vector embeddings to cluster similar documents together in a low-dimensional vector space.…”

“…The Springer dataset used in this work is the one used also in [11]. It is a collection of 56,405 multi-disciplinary e-book titles from Springer, subscribed to by the Hellenic Academic Libraries Link (HEAL-Link, https://www.heallink.gr/en/home-2/).…”

“…The metadata of each book in the collection includes information regarding the title, subtitle, authors, publication year, publisher, table of contents (ToC) and subject. Similarly, to [11], we chose to model each book by its ToC for a number of good reasons explained in the Introduction. The categorization of the collection according to Springer's main subjects is outlined in Table II.…”

“…As one can observe from Table II, most documents belong to only 5 categories (Computer Science, Physical Sciences, Engineering, Medicine and Mathematics) and the distribution among the classes is not equal. This fact poses challenges concerning the quality measures that are often used to measure the performance of a classifier, as the accuracy on the smaller categories may deteriorate over the more significant categories, as explained in [11]. At first, the dataset had 252 different categories, but those that could be grouped together were mapped to more general ones, so there were only 26 general categories in the end.…”

BERT-Based Hybrid RNN Model for Multi-class Text Classification to Study the Effect of Pre-trained Word Embeddings

“…MOOC, IMDB-2, IMDB-10 and SST-5. The CNN+LSTM architecture was also used by Giannopoulou et al [7] to categorize e-books into pre-defined book categories using their table of contents as the text samples. Software As-A Service (SaaS) is one of the popular software delivery models.…”