Representation of texts as complex networks: a mesoscopic approach

Abstract

Statistical techniques that analyze texts, referred to as text analytics, have departed from the use of simple word count statistics towards a new paradigm. Text mining now hinges on a more sophisticated set of methods, including the representations in terms of complex networks. While well-established word-adjacency (co-occurrence) methods successfully grasp syntactical features of written texts, they are unable to represent important aspects of textual data, such as its topical structure, i.e. the sequence of subjects developing at a mesoscopic level along the text. Such aspects are often overlooked by current methodologies. In order to grasp the mesoscopic characteristics of semantical content in written texts, we devised a network model which is able to analyze documents in a multi-scale fashion. In the proposed model, a limited amount of adjacent paragraphs are represented as nodes, which are connected whenever they share a minimum semantical content. To illustrate the capabilities of our model, we present, as a case example, a qualitative analysis of "Alice's Adventures in Wonderland". We show that the mesoscopic structure of a document, modeled as a network, reveals many semantic traits of texts. Such an approach paves the way to a myriad of semantic-based applications. In addition, our approach is illustrated in a machine learning context, in which texts are classified among real texts and randomized instances.

On the "Calligraphy" of Books

“…Those networks, for example, are unable to represent the topical structure found in many texts. So as to overcome such a limitation, a mesoscopic representation has been recently proposed (de Arruda et al, 2017). The main goal of that approach was to take into account the semantical relationship between chunks of text.…”

“…There are several ways to represent texts as complex networks, such as co-occurrence, syntactic, semantic or similarity networks (Mihalcea and Radev, 2011;Cong and Liu, 2014). In this study, we adopt the mesoscopic network approach proposed by de Arruda et al (2017). Such networks are able to represent the text unfolding along time, which is normally overlooked by traditional approaches.…”

“…Co-occurrence networks do not portray the topical structure found in many texts and are usually devoid of community structure (de Arruda et al, 2016). In order to overcome this disadvantage, some techniques have been devoted to the mesoscopic representation of texts (de Arruda et al, 2016(de Arruda et al, , 2017. de Arruda et al (2017) proposed a novel networked model, in which each node represents a respective set of consecutive paragraphs, while weighted edges express the similarity between nodes.…”

“…In order to overcome this disadvantage, some techniques have been devoted to the mesoscopic representation of texts (de Arruda et al, 2016(de Arruda et al, , 2017. de Arruda et al (2017) proposed a novel networked model, in which each node represents a respective set of consecutive paragraphs, while weighted edges express the similarity between nodes. Their proposed network is able to extract the organization and flow of text by effectively capturing the similarity between the blocks of text.…”

Image Converting into Complex Networks : Scale- Level Segmentation Approach

“…The paper [26] elaborated a network model which analyzes documents within a multi-scale context . In the proposed model a mesoscopic representation of texts was considered, where a node represented a large context, e.g.…”