Business Intelligence Investigate

Predicting Sequential Pattern Changes in Data Streams

Download now Free registration required

Executive Summary

Data streams are utilized in an increasing number of real-time information technology applications. Unlike traditional datasets, data streams are temporally ordered, fast changing and massive. Due to their tremendous volume, performing multiple scans of the entire data stream is impractical. Thus, traditional sequential pattern mining algorithms cannot be applied. Accordingly, the present study proposes a new sequential pattern mining model for predicting sequential pattern changes in data streams. The experimental results show that the prediction performance of the proposed model is better than that of two linear regression-based models.

  • Format: PDF
  • Size: 500.15 KB