RETRACTED ARTICLE: Applying data mining techniques to explore user behaviors and watching video patterns in converged...

Area of Science:

• Human-Computer Interaction
• Data Mining
• Multimedia Systems

Background:

• Web multimedia systems facilitate diverse user interactions for leisure and entertainment.
• Information technology (IT) systems enhance multimedia engagement and user interactions.
• Understanding user behavior in multimedia consumption is crucial for system optimization.

Purpose of the Study:

• To analyze user behaviors in viewing multimedia videos by key points in time.
• To explore user watching patterns in converged IT environments using data mining.
• To categorize user engagement levels based on video player interaction data.

Main Methods:

• Data mining techniques were applied to analyze user video watching patterns.
• System logs from a web multimedia video player were collected and classified.
• K-means clustering was used to group users into three engagement categories based on four variables: playing time, active playing time, played amount, and actively played amount.

Main Results:

• User watching patterns were successfully clustered into three distinct categories: actively engaged, watching engaged, and long engaged users.
• Significant differences were found in the watching behaviors among the three identified user categories.
• The analysis provided insights into how users interact with web multimedia content.

Conclusions:

• User engagement with multimedia content varies significantly and can be categorized.
• Data mining and clustering techniques are effective in identifying and understanding diverse user behaviors in multimedia systems.
• These findings can inform the design and improvement of web multimedia platforms to better cater to different user engagement styles.