Uday V. Kulkarni

Hybrid personalized recommender system using centering-bunching based clustering algorithm

In recent years, there is overload of products information on world wide web. A personalized recommendation is an enabling mechanism to overcome information overload occurred when shopping in an Internet marketplace. This paper proposes a novel centering-bunching based clustering (CBBC) algorithm which is used for hybrid personalized recommender system (CBBCHPRS). The proposed system works in two phases. In the first phase, opinions from the users are collected in the form of user-item rating matrix. They are clustered offline using CBBC into predetermined number clusters and stored in a database for future recommendation. In the second phase, the recommendations are generated online for active user using similarity measures by choosing the clusters with good quality rating. This helps to get further effectiveness and quality of recommendations for the active users. The experimental results using Iris dataset show that the proposed CBBC performs better than K-means and new K-medodis algorithms. The performance of CBBCHPRS is evaluated using Jester database available on website of California University, Berkeley and compared with ants recommender system (ARS). The results obtained empirically demonstrate that the proposed CBBCHPRS performs superiorly and alleviates problems such as cold-start, first-rater and sparsity.

Fuzzy hyperline segment neural network for rotation invariant handwritten character recognition

In this paper fuzzy hyperline segment neural network (FHLSNN) is proposed which is used for recognition of handwritten characters. The FHLSNN utilizes fuzzy sets as pattern classes in which each fuzzy set is an union of fuzzy set hyperline segments. The fuzzy set hyperline segment is an n-dimensional hyperline segment defined by two end points with a corresponding membership function. The handwritten characters can be in arbitrary location, scale and orientation. After moment normalization rotation invariant ring-data and Zernike moment feature vectors are extracted from characters. Finally, FHLSNN algorithm is used to classify these feature vectors by its strong ability of discriminating ill-defined character classes. The FHLSNN algorithm is compared with fuzzy neural network proposed by Kwan and Cai (1994), which is modified to work under supervised environment and fuzzy min-max neural network proposed by Simpson (1992, 1993). The FHLSNN algorithm is found to be superior with respect to the training time, recall time per pattern and the generalization.

A New Approach for on Line Recommender System in Web Usage Mining

The Internet is one of the fastest growing areas of intelligence gathering. During their navigation Web users leave many records of their activity. This huge amount of data can be a useful source of knowledge. Sophisticated mining processes are needed for this knowledge to be extracted, understood and used. Web Usage Mining (WUM) systems are specifically designed to carry out this task by analyzing the data representing usage data about a particular Web Site. WUM can model user behavior and, therefore, to forecast their future movements. Online prediction is one Web usage mining application. However, the accuracy of the prediction and classification in the current architecture of predicting userspsila future requests systems can not still satisfy users especially in huge Web sites. To provide online prediction efficiently, we develop architecture for online recommendation for predicting in Web Usage Mining System .In this paper we propose architecture of on line recommendation in Web usage mining (OLRWMS) for enhancing accuracy of classification by interaction between classifications, evaluation, and current user activates and user profile in online phase of this architecture.

General fuzzy hyperline segment neural network

We propose a general fuzzy hyperline segment neural network (GFHLSNN) and its learning algorithm, which is an extension of the fuzzy hyperline segment neural network proposed by Kulkarni et al (2001). It combines supervised and unsupervised learning in a single algorithm so that it can be used for pure classification, pure clustering and hybrid classification/clustering.

Fuzzy hypersphere neural network classifier

In this paper fuzzy hypersphere neural network (FHSNN) is proposed with its learning algorithm. The FHSNN utilizes fuzzy sets as pattern classes in which each fuzzy set is an union of fuzzy set hyperspheres. Its performance is compared with other two fuzzy neural networks and found to be superior with respect to the training time, recall time per pattern and the generalization.

Offline Handwritten Character Recognition of Gujrati script using Pattern Matching

during the last forty years, Handwritten Character Recognition (HCR) has most often been investigated under the framework of Character Recognition (OCR) and Pattern Recognition. HCR is more considered as a perceptual and interpretation task closely connected with research into Human Language. India is a country which uses many languages in the different parts of the country be it for personal use or use of business. In this study we propose a novel solution for performing character recognition in Gujrati, the official language of Gujarat. Pursued by the preprocessing techniques, we suggest a method called Pattern Matching where a character is identified by analyzing its shape and comparing its features that distinguish each character. Various handwritten characters from forms or peripheral devices etc. are recognized with the help of various pre-processing and image enhancement techniques. These characters are further more specifically recognized by Pattern matching using Neural Network.

Trends in Handwriting Recognition

The field of Handwriting recognition has evolved over the past three or four decades into a broad based activity which has had a measurable impact on applications. Some of the most significant practical impact has occurred in the past decade in handwriting recognition. Successful application of the established methods requires good understanding of their behavior and how well they match a particular context. Difficulties can arise from either the intrinsic complexity of a problem or a mismatch of methods to problems. Many emerging applications of involve complicated high-dimensional pattern spaces, small amounts of data-per-dimension, low signal-to-noise ratio, poorly specified statistical distributions, and anomalous statistical outliers. In some cases these difficulties are compounded by distributed data collection requirements that impose constraints on data integration and decentralized decision making. This creates both challenges and opportunities for Handwriting recognition research. This survey divides various approaches to handwriting recognition in nine different categories. Authors explore resent trends in Handwriting recognition and describe the areas of challenges and some possible solutions.

Speech recognition using modified fuzzy hypersphere neural network

In this paper, a modified fuzzy hypersphere neural network (MFHSNN) is proposed, which is an extension of the fuzzy hypersphere neural network (FHSNN) proposed by Kulkarni and Sontakke (2001). Its performance is compared with FHSNN for the recognition of spoken Marathi (the language spoken in the state of Maharashtra, India) digits and found to be superior with respect to the recall time and recognition rate.

General fuzzy hypersphere neural network

This paper describes a general fuzzy hypersphere neural network that uses supervised and unsupervised learning within a single training algorithm. It is an extension of fuzzy hypersphere neural network and can be used for pure classification, pure clustering or hybrid clustering/classification.

Template Matching Algorithm for Gujrati Character Recognition

at the dawn of the 3rd millennium, Human Handwriting Recognition is emerging from its infancy and set to become a mature technique. We shall probably see in the near future a number of mixed systems able to read both online and off-line handwriting. In this study we propose a simple yet robust structural solution for performing character recognition in Gujrati, the official language of Gujarat. Pursued by the preprocessing techniques, we suggest a method called template matching where a character is identified by analyzing its shape and comparing its features that distinguish each character. The algorithm appears to be very robust against stroke order variations and large shape variations. The results seem encouraging.