Detecting Anomalies in Data Stream Using Efficient Techniques: A Review
Venisha Maria Tellis, Divya Jennifer Dsouza · 2018 International Conference on Control, Power, Communication and Computing Technologies (ICCPCCT) · 2018
As data is a critical part of many application data mining is a most researched domain in the today's world. By discovering normal tendency or dealing out of data usually we find information but infrequent occurrence or data item sometimes may provide details which is very useful to us. One of the function of data mining is anomaly detection, it finds unusual data or order not visible in the dataset. It is also one of the important problem in data mining. Unlimited order of data with direct or indirect material conditions is called as data stream, it is dynamic and unknown in nature. For static data, that requires full dataset for modelling, traditional anomaly detection techniques does not fit for data stream because full data stream can't be stored. In detection of unknown data, in certain applications like calling cards, criminal behaviors, finding computer intrusion, detection of fraud in credit cards etc. finding anomalies is important. In this survey, for finding anomaly in data stream usual methods or techniques have been described.