Behavioral Data Modeling of Language Games via Time Series Forecasting and Clustering Techniques

View at publisher Summarise with VIRAT
Authors: Xiaochun Zhang, Xu An, Ruixin Xue
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

With the wide dissemination of interactive language games on online platforms, modeling and analysis based on user behavior data have increasingly become an important research direction in the fields of natural language processing and intelligent prediction. This paper proposes a modeling framework that integrates time series prediction and cluster analysis for the performance reporting behavior and word difficulty structure of players in online language games. In the time series modeling section, the ARIMA and SARIMA models were respectively constructed to fit and predict the number of daily user reports, and comparisons were made from aspects such as goodness of fit, error indicators, and residual analysis. The experimental results show that the SARIMA model outperforms the ARIMA model in capturing periodic trends and improving prediction accuracy. In the cluster analysis section, this paper extracts the correct guess distribution of words in the six-round word guessing as the feature variable, and uses the K-means algorithm to perform unsupervised classification of word difficulty. The clustering effect is significant, and there are highly significant differences in the dimension of word guessing behavior among various types. Overall, the dual-model framework constructed in this paper not only enhances the accuracy of modeling user behavior but also provides a theoretical basis and algorithmic support for the content optimization and intelligent recommendation of language-based games.

Fields of Study & Subjects

Time Series Analysis and Forecasting Data Mining and Machine Learning Applications Online Learning and Analytics

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
291–296
Publication date
28 January 2026
Language
EN