Analysis of the Influence of Text Features on the Usefulness of Information: A Case of Tourism Text

Abstract

This paper takes 9260 domestic free tourism strategy data collected from the Mafengwo website as samples, quantifies the travel guide texts by using data mining, and performs word frequency statistics and keyword extraction on the data through Python code and NLPIR platform, then divides the high-weight words into three categories that affect the usefulness of information through hierarchical clustering. Ten hypotheses are proposed in the paper based on previous research. And a negative binomial regression model is built to conduct analysis. The results show that when the number of reads is regarded as a control variable, all ten text features, such as the rate of containing pictures, have a significant positive relationship on information usefulness. Therefore, suggestions are provided to develop influential and high-quality tourism strategies in terms of text features.