A novel prediction algorithm for multivariate data sets

Regression analysis is a statistical technique that most commonly used for forecasting. Data sets are becoming very large due to continuous transactions in today's high-paced world. The data difficult manage and interpret. All the independent variables can’t be considered prediction because it costs high maintenance of set. A novel algorithm has been implemented this paper. Its emphasis on extraction efficient from various selection based Mean Square Errors (MSE) as well coefficient determination r2p, after final equation framed basis deviation actual mean. This which evaluate four parameters: Root Error (RMSE), Absolute (MAE), Percentage (MAPE) residuals. multivariate set with low costs, preprocessing lower root mean square error For one dimensional, two-dimensional, frequent stream data, time series proposed can also used. impact enhance accuracy rate forecasting minimized average rate.