Abstract Data mining is concerned with the analysis of data for finding patterns and regularities in the data sets. Statistics is a mathematical science concerned with the collection, analysis, interpretation or explanation, and presentation of data. Statistics plays a very important role in the process of data mining analysis and equally visualization of data plays a very important role in decision making process. Instance Based Learning Streams is an instance-based learning algorithm used to perform regression analysis on data streams. The algorithm is able to handle large data streams with less memory and computational power. The paper aims at the implementation of Instance Based Learning Streams as an extension to the massive online analysis framework for data stream mining to develop a regression model. The study reveals that the regression analysis could be performed not only on small data sets but also on data streams as in the present case but the method of analysis will be different in the two cases. In the case of small data set the regression models are linear, multiple and polynomial, while in the case of data streams the entire analysis is performed under the massive online analysis framework by taking the two evaluation parameters basic regression performance evaluator and windows regression performance evaluator. This finding is first of its kind in literature. *Author for correspondence Indian Journal of Science and Technology, Vol 7(6), 864–870, June 2014 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 Regression Model using Instance based Learning Streams P. K. Srimani 1 and Malini M. Patil 2* 1 R&D Division, Bangalore University, Bangalore, Karnataka, India; profsrimanipk@gmail.com 2 Department of ISE, J.S.S. Academy of Technical Education, Bangalore, Karnataka, India; patilmalini31@yahoo.com Keywords: Data Streams, IBLStreams, Instance, Prediction, Massive Online Analysis, Regression 1. Introduction Advancement of technology has resulted in large stor- age of data. hese large masses of data consist of some hidden information of strategic importance, which can be used for future analysis with efective decision mak- ing. he two important types of data analysis methods are classiication and prediction. In the former case a model is constructed (classiier) to predict the categorical labels, in the latter case a model is constructed (predictor) to predict the continuous variables. Regression analysis is a statistical method which is widely used for prediction and forecasting. It is also considered as the part of machine learning process. It plays a very important role in pre- diction. he authors 1 have explained thoroughly about regression analysis. It can be used to model the relation- ship between one or more independent variables and a dependent variable. Diferent techniques are developed to carry out regression analysis, namely, Linear regression, multiple regression, polynomial regression and ordinary least squares regression. hese methods can be parametric or non parametric in nature. Many tools are also developed to carry out regression analysis viz. miniTab, Gnumeric, PASW. Regression analysis can be performed very ei- ciently using MS-Excel also. Rapid advancement of the technology resulted in the storage of digital data has also increased very rapidly. he continuous arrival of data is referred to as data stream. Network monitoring data, sensor data, web clicks, usage of credit cards weather forecasting data are few exam- ples of data streams. he data streams are massive in nature and they arrive at a very high speed. Data min- ing techniques are not suitable for mining data streams and the data streams must be processed under very strict constraints of space and time. Gaber et al. 2 , Gaber and