Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/195368
Title: | Design and Implementation of An Agent Based Stock Clustering for An Efficient Portfolio Management |
Researcher: | Dalal, Preeti |
Guide(s): | Saini, Jitendra |
Keywords: | Dunn Index Earnings per Share(EPS) Fast k-Means F-Measure Intra-class inertia k-Means kMedoids Portfolio Management Price per Earnings (P/E) Silhouette index |
University: | RK University |
Completed Date: | 16/01/2017 |
Abstract: | The cluster analysis plays an important role in various financial domains. It is a process of creating groups by assigning data sets in different groups. The characteristic of data sets in same group is similar than data sets in other groups. It is an unsupervised technique in which the class labels of datasets are not previously known. This research work proposed a framework for building an efficient portfolio management using clustering of Nifty financial data through intelligent agents. The framework uses various kinds of agents, i.e. user agent, clustering agent, ranking agent, validation agent and portfolio manager automate portfolio management task. The user agent interacts with users and accepts user investment. Then ranking agent assigns weights to attributes and higher weighted attributes are selected for data mining task. Clustering agents for each clustering algorithm helped to detect clusters automatically from financial data by sending request to data agents which collect necessary financial data required for clustering task. This research work used only centroid based clustering algorithms. The clustering agents sent cluster result to validation agent which validates clustering results. The validity of clustering result is based on similarity and dissimilarity measure. newlineVII Preeti B. Dalal newlineABSTRACT newlineThe performance analysis of three clustering algorithm was evaluated using intra-class inertia which finds best clustering algorithm. Then portfolio manager constructed diversified portfolio of generated clusters and suggests optimum portfolio for investment which gives maximum possible return at the given level of risk. An investor can invest money in different securities instead of investing in a single security so that he/she can get a maximum return and a minimum risk. This agent-based framework assists investors to achieve better investment decision. This research study gives an innovative approach to formulate an efficient portfolio. newline |
Pagination: | |
URI: | http://hdl.handle.net/10603/195368 |
Appears in Departments: | Faculty of Technology |
Files in This Item:
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: