Complex networks analysis of language complexity

Abstract

-Methods from statistical physics, such as those involving complex networks, have been increasingly used in quantitative analysis of linguistic phenomena. In this paper, we represented pieces of text with different levels of simplification in co-occurrence networks and found that topological regularity correlated negatively with textual complexity. Furthermore, in less complex texts the distance between concepts, represented as nodes, tended to decrease. The complex networks metrics were treated with multivariate pattern recognition techniques, which allowed us to distinguish between original texts and their simplified versions. For each original text, two simplified versions were generated manually with increasing number of simplification operations. As expected, distinction was easier for the strongly simplified versions, where the most relevant metrics were node strength, shortest paths and diversity. Also, the discrimination of complex texts was improved with higher hierarchical network metrics, thus pointing to the usefulness of considering wider contexts around the concepts. Though the accuracy rate in the distinction was not as high as in methods using deep linguistic knowledge, the complex network approach is still useful for a rapid screening of texts whenever assessing complexity is essential to guarantee accessibility to readers with limited reading ability.

Comparison of the language networks from literature and blogs

“…Liu and Cong [14] constructed co-occurrence networks from text in different languages and used complex network parameters for the classification (hierarchical clustering) of 14 languages, where Croatian was amongst 12 Slavic. Different applications of linguistic network analysis in NLP includes: evaluation of language complexity [15], automatic summarisation [16] and evaluation of machine translation [17], authorship attribution [18] and text quality analysis [19].…”

“…Costa et al [15] studied the relationship between the topology of network and complexity of the text. They studied texts with different levels of simplification in cooccurrence networks and found that topological regularity correlated negatively with textual complexity.…”

“…Various types of linguistic networks have already been studied: syntax networks [1,2], semantic networks [3], phonological networks [4], syllable networks [5,6], word co-occurrence networks [7][8][9][10][11][12][13][14][15][16][17][18][19]. In [3,20,21] a systematic methodological overview of linguistic complex networks principles is presented.…”

“…In [3,20,21] a systematic methodological overview of linguistic complex networks principles is presented. Recently, linguistic co-occurrence networks have been intensively studied in order to analyse the structure of the language [7][8][9][10][11][12][13][14][15][16][17][18][19].…”

Using complex networks for text classification: Discriminating informative and imaginative documents

“…Of special relevance to this paper, are the networked models of language and texts, which have been useful to unveil universal properties including the scale-free and small-world phenomena [22]. In practical terms, networked models have been useful to grasp several features of texts, such as quality [23], complexity [24] and authenticity [25]. Particularly, in this study, we used the so-called word adjacency model, which is a approximation of text networks formed by syntactical links [26].…”