International Journal of u- and e- Service, Science and Technology Vol. 6, No. 4, August, 2013 155 SVM Regression and SONN based approach for seasonal crop price prediction K.Lavanya 1 , T.Raguchander 2 and N.Ch.S.N.Iyengar 3 1,3 School of Computing Science and Engineering, VIT University, Vellore , India 2 Tamilnadu Agricultural University, Coimbatore, India lavanya.k@vit.ac.in, raguchander@rediffmail.com, nchsniyengar48@gmail.com Abstract In Agriculture pre harvest glut, post harvest loss and intermediary involvement cause producer to get the lowest price in the entire marketing process. Though the authorities provide long time monthly price series for various geographically spread regions of the country in websites, the lack of knowledge to process online market data does not allow farmers to take a dynamic decision for the sale of their produce. In this model, the equipped marketing information is analyzed for integrating domestic markets and manifesting price transmission from markets to farm gate (prices received by farmers) leading to sustainable profit. Market wise support price for crops is extracted from website tables using XML SAX(Simple API for XML) Expression and interested crop price associated with its season and variety are clustered using SONN (Self organized neural network). Then support vector regression (SVR) based price prediction for long and midterm movement helps farmers in the decision making of agricultural marketing. As all the processing is done at the back end it is easy for the farmers to interact with the system and reach for more accurate prediction. The results assure of more precision than other traditional methods such as regression techniques and moving averages Keywords: XML SAX Expression, price extraction, SONN, clustering, SVR, prediction, sustainable profit 1. Introduction Indian Agricultural Market structure is a network of cooperatives at the local, regional, state and national levels. Over a period of time, these markets have acquired the status of limiting and monopolistic, providing no help in direct and free marketing, thereby preventing private investments in the sector. The misuse by intermediaries, exporters, processors and traders prevents farmers from getting premium price and timely payment for their produce [4]. Nowadays, finding a suitable market for the marketed surplus with the increased agricultural production is a big challenge. For example, if crop production during the season is surplus, it ends up in post harvest loss of product. But if the same product is transported to another market where there is a need and preference, considerable profit could be achieved. The changing global agricultural scenario insists on review of the policies related to pricing, marketing and trading of agricultural commodities. In order to provide dynamism and efficiency in the marketing system online market wise monthly series price data were made available for direct usage by farmers. This amplified marketing information has a positive effect in decision making, but collecting and disseminating information has been often complicated and expensive. Data mining techniques have a great deal of interest for price forecasting in recent years that aids huge amount of data processing for better price realization [3, 8].