Showing papers by "All Saints' College published in 2010"

SOM and Technical Indicators Based Hybrid Model Gives Better Returns on Investments as Compared to BSE-30 Index

Abstract: Selection of stocks that are suitable for investment is a complex task. The main aim of every investor is to earn maximum possible returns on investments. There are many conventional techniques being used and these include technical and fundamental analysis. The main issue with any approach is the proper weighting of criteria to obtain a list of stocks that are suitable for investments. This paper proposes an improved method for stock picking using self-organizing maps and technical analysis. The stock selected using this method outperformed the BSE-30 Index by about 28.41% based on one month of stock data.

ART and Modular Neural Network Architecture for Multilevel Categorization and Recognition of Fingerprints

Abstract: ART1 based clustering approach is used for classification, which groups fingerprints into more compact classes. ART1 is a efficient technique for grouping fingerprints in to N number of classes, which speedup the process of fingerprint recognition. After classification of fingerprints the key-fingerprint class is used for the purpose of fingerprint identification. The key-fingerprint is recognized by using Monolithic and Modular Neural Network and their performance has been compared on the bases of time and accuracy. Due to modularity, Modular Neural Network gives better performance on the classified databases as compared to Monolithic Neural Network even with poor quality fingerprints. Monolithic Neural Network takes average of 44.7 seconds with an accuracy of 98%, correct recognition where as Modular Neural Network takes average time 1.84 seconds with an accuracy of 100% correct recognition