Ali Ilkhani - Department of Knowledge Management Tehran Securities Exchange Technology Management Co, Tehran, Iran
Golnoosh Abaee - Department of Computer Science Islamic Azad University Roodehen Branch

Time series data poses a significant variation to the traditional segmentation techniques of data miningbecause the observation is derived from multiple instances of the same underlying record. In this paperwe propose a new pattern for extracting knowledge form stock market by eliminating some partialfluctuation and using clustering algorithm of data mining which bring us efficient information aboutcurrent situation. Since similarity measurements of time series play a crucial role in many KDDapplications, data mining clustering techniques could be used in extracting hidden information at equaltime intervals. Evaluating and analyzing the results help us; find efficient reasons on stock marketingportfolios variation.

Pattern, Clustering, Data Mining, Similarity, KDD, Time Series