1. UKM Learning and Research Repository
2. PHD Thesis / Tesis PHD
3. Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Title:	Financial pattern discovery using similarity mining technique
Authors:	Abd El Salam Tamer Hassan (P58744)
Supervisor:	Zalinda Othman, Assoc. Prof. Dr.
Keywords:	Dissertations, Academic -- Malaysia
Issue Date:	22-Sep-2014
Description:	Time series data representation is a fundamental problem assimilated in the high dimensionality of time-series data. Dimensional reduction strategy is one of the principal requirements for a successful representation to improve the efficiency of extracting the attractive trend patterns on the time series data. Furthermore, an efficient and accurate similarity searching on a huge time series data set is a crucial problem in data mining pre-processing. Specifically in financial time series, another problem arises due to the lack of prior work analyzing the effect of the online investor sentiment on the characterization of the financial fluctuations. Symbolic representations have proven to be a very effective way to reduce the dimensionality of time series without loss of knowledge. However, symbolic representations suffer from another challenges promoted by the possibility of losing some principal patterns due to the difficulty of properly determining the reduction scale of the desired representation combined with the impractical utilization of dealing with the whole data with the same weight. The main objective of this thesis is to integrate Symbolic Aggregate Approximate and Piecewise Linear Approximation techniques in searching a pattern similarity in stock market index and investor sentiment. This integration is proposed to overcome symbolic representation pattern mismatch. Moreover, the data dimensionality is reduced by keeping more detail on recent-pattern data and less detail on older ones using modified sliding window. The modified sliding window is controlled by the corresponding classification error rate. Experimental tests were made on the UCR standard dataset comparing with the state of the art techniques such as Discrete Fourier Transform (DFT) and Dynamic Time Warping (DTW). The proposed techniques showed promising results. Furthermore, practical experiments were made on the Egyptian stock market indices EGX30, EGX70 and EGX100 and sentiment data produced by news and investor reviews. The discovered patterns were reviewed by professional financial experts. The results showed the accuracy and effectiveness of the proposed approach with classification error rate of 18%.,Ph.D
Pages:	175
Call Number:	QA76.9.D343 A234 2014
Publisher:	UKM, Bangi
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.