With the increasing availability of passive, wearable sensor devices, digital lifelogs can now be captured for individuals. Lifelogs contain a digital trace of a person's life, and are characterised by large quantities of rich contextual data. In this paper, we propose a content-based recommender system to leverage such lifelogs to suggest activities to users. We model lifelogs as timelines of chronological sequences of activity objects, and describe a recommendation framework in which a two-level distance metric is proposed to measure the similarity between current and past timelines. An initial evaluation of our activity recommender performed using a real-world lifelog dataset demonstrates the utility of our approach.

No abstract

Data mining has been used extensively and broadly by several network organizations. Classification based algorithms provide a significant advantage in order to detect attacks in the training data. Network applications usage is being increased every day as the internet usage is exponentially increasing. In the same way, Network attacks detection is gradually decreased as data source is increasing. There is a need to develop some robust decision tree in order to produce effective decision rules from the attacked data. In this paper improved, decision tree is implemented in order to detect network attacks like TCP SYN , Ping of Death, ARP Spoof attacks. This improved tree is also tested on famous network intrusion dataset Kddcup 99 dataset. Experimental result shows this improved decision tree classifier gives effective decision rules compare to existing decision tree techniques like ID3 and C45 algorithms. Finally, this robust decision tree evaluates less false positive and true negative alarm rates compare to existing algorithms.