A stochastic model of word occurrences in hierarchically structured written texts

• DOI: 10.1007/s42452-022-04953-w
• Publication Date: 2022-02-14T10:02:51Z
• Authors: Hiroshi Ogura; Yasutaka Hanada; Hiromi Amano; Masato Kondo
• Keywords: Technology; Word occurrence; Long-range correlation; Science; T; Q; Autocorrelation function; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology; Waiting time distribution; Stochastic process

Abstract In previous studies, we have treated real written texts as time series data and tried to investigate dynamic correlations of word occurrences by utilizing autocorrelation functions (ACFs) also simulation pseudo-text synthesis. The results showed that words appear in can be classified into two groups: a group showing (Type-I words), no (Type-II words). this study, the characteristics these types terms their waiting distributions (WTDs) occurrences. for Type-II show stochastic processes govern generating are superpositions Poisson point with various rate constants. We further propose model WTDs Type-I which hierarchical structure is considered. agree well predictions proposed model, indicating important long-range words.