Interest Portrait Construction and Dynamic Evolution Analysis Based on Multi-Source Behavioral Data

DOI: https://doi.org/10.62517/jike.202604217

Author(s)

Shuaizhe Huang

Affiliation(s)

School of Mathematics and Statistics, Anqing Normal University, Anqing, Anhui, China *Corresponding Author

Abstract

In the digital era, diverse and heterogeneous user behavior data are trapped in data silos. Traditional static or single-source interest profiling methods cannot support precise user modeling. This paper focuses on constructing user interest portraits and analyzing their dynamic evolution using multi-source heterogeneous behavioral data. It integrates e-commerce behavioral logs (from the Taobao platform) and movie rating interaction data (from the MovieLens dataset), constructs a unified behavior representation via data cleaning, alignment, and fusion, uses NLP to extract fine-grained interest tags, and introduces a time-series framework with a sliding window and attention mechanism to capture dynamic features and identify persistence, mutation, and periodic patterns of user interests.

Keywords

User Interest Profile; Multi-source Heterogeneous Data; Data Fusion; Dynamic Evolution; Natural Language Processing(NLP)

References

[1] Tengfei Ren, Zhipeng Zhang, Bo Jia, Shiwen Zhang. Retrieval-Augmented Generation-aided causal identification of aviation accidents: A large language model methodology[J]. Expert Systems With Applications,2025,278 [2] Moritz Lell, Abhishek Gogna, Vincent Kloesgen, Ulrike Avenhaus, Jost Dörnte, Wera Maria Eckhoff, Tobias Eschholz, Mario Gils, Martin Kirchhoff, Michael Koch, Sonja Kollers, Nina Pfeiffer, Matthias Rapp, Valentin Wimmer, Markus Wolf, Jochen Reif, Yusheng Zhao. Breaking down data silos across companies to train genome-wide predictions: A feasibility study in wheat [J]. Plant biotechnology journal,2025,23(7): [3] Bingchen Li, Changdong Li, Yong Liu, Jie Tan, Pengfei Feng, Wenmin Yao. Harnessing Distributed Deep Learning for Landslide Displacement Prediction: A Multi-Model Collaborative Approach Amidst Data Silos[J]. Journal of Earth Science,2024,35(5): [4] Zachary F Fisher, Younghoon Kim, Vladas Pipiras, Christopher Crawford, Daniel J Petrie, Michael D Hunter, Charles F Geier. Structured Estimation of Heterogeneous Time Series [J]. Multivariate behavioral research,2024,59(6): [5] Longyun Chen, Chen Qiao, Kai Ren,Gang Qu, Vince D Calhoun, Julia M Stephen, Tony W Wilson, Yu Ping Wang. Explainable spatio-temporal graph evolution learning with applications to dynamic brain network analysis during development [J]. NeuroImage,2024,298 [6] Luise von Keyserlingk, Fani Lauermann, Qiujie Li, Renzhe Yu, Charlott Rubach,Richard Arum,Jutta Heckhausen. Students' study activities before and after exam deadlines as predictors of performance in STEM courses: A multi-source data analysis[J]. Learning and Individual Differences,2025,117 [7] Shichun Z, Wang L, Xing Z. A Study on Visitors’ Tour Patterns and Perceptual Experiences in VR Exhibitions Through the Integration of Multi-Source Behavioral Data [J]. International Journal of Human–ComputerInteraction, 2025, 41 (24): 15893-15914. [8] Dequan Zhang, Wei Jiang, Jincheng Lou, Xuanzhou Han, Jianye Xia. Biofuser: a multi-source data fusion platform for fusing the data of fermentation process devices[J]. Frontiers in Digital Health,2024,6 [9] Jianhui Xu, Mustafa Man, Ily Amalina Ahmad Sabri, Guoyi Li, Chao Yang, Mingxue Jin. Research on Personalized Recommendation of High-Quality Academic Resources based on user Portrait[J]. International Journal of Advanced Computer Science and Applications (IJACSA),2022,13(10): [10] Fan Hongmin, Liu Chang, Wang Qing song. Measurement, Spatial Differences, and the Dynamic Evolution of China's Urban Business Environment Levels[J]. Journal of Urban Planning and Development,2024,150(2): [11] Alibaba Group. User Behavior Data from Taobao for Recommendation[DS/OL]. Tianchi Big Data Dataset, 2018. https://tianchi.aliyun.com/dataset/649. [12] Harper, F. M., & Konstan, J. A. (2015). The MovieLens Datasets: History and Context. ACM Transactions on Interactive Intelligent Systems, 5(4), 1-19. [13] Bobzien L, Verwiebe R, Kalleitner F. Visualizing age-specific digital platform usage in Germany[J]. Socius: Sociological Research for a Dynamic World, 2025, 11: 1-4. https://journals.sagepub.com/doi/10.1177/23780231251319360 [14] Ding Hao, Li Shuqing. Personalized Recommendation Based on Predictive Analysis of User’s Interests[J]. Data Analysis and Knowledge Discovery, 2019, 3(11): 43-51 https://doi.org/10.11925/infotech.2096-3467.2019.0370