Recognition method of behavioral pattern by sequential learning for life logs

We proposed a new pattern recognition method suitable for sequential learning and self adapting. which meets the requirement of Web behavioral analysis on large Life Log databases. Even if sparse data, such as web page keywords, a user´s web sequence can be efficiently compared with the statistical sequence. If a user´s Web behavior does not change suddenly, our method will be accurate. For future work, we will evaluate the above metioned details in terms of performance and accuracy.