Data Mining for Cotton Yield Prediction in North Region of Gujarat India

Abstract

Data Mining analysis is conducted in order to discover the new knowledge in the form of newlinepatterns, trends, and associations, that are not achieved through statistical tools. The newlineamount of data gathered and stored on many fronts has grown tremendously and the newlineprocess of knowledge discovery (data mining) from these data has become, complex newlineand very important for the government, businesses, science and research communities. newlineThis research used time series analysis using Gaussian processes over relational data newlinethat enables the declarative formulation, systematic optimization, and processing of a newlinelarge class of mining queries. Time series analysis is the process of using statistical newlinetechniques to model and explain a time-dependent series of data points. Time series newlineforecasting is the process of using a model to generate predictions (forecasts) for future newlineevents based on known past events. newlineCotton is a very important crop, as India leads it in terms of production in the world; and newlinealso that a vast number of manpower is engaged in farming as well as post harvest newlineprocessing and management of different derivatives of it. Weather is crucial for the newlineproductivity of the crop. All the talukas of the three districts of North Gujarat where newlinecotton is cultivated have been selected purposively for this study. The effect of soil type, newlinesoil pH, soil organic carbon, phosphorous, potassium, precipitation and temperature newlinewere selected as independent factors. The weather data for this study was collected newlinefrom Sardarkrushinagar Dantiwada Agricultural University; Dantiwada and soil newlineparameter data for the selected study area was collected from Anand Agricultural newlineUniversity; Anand. The data of cotton crop yield for the ten years (2006 to 2015) were newlineobtained from the Department of Agriculture, Government of Gujarat. newlineThe WEKA algorithms such as Multilayer Perceptron, SMOreg, Kster, Additive newlineRegression, and Gaussian Process were used for finding the correlation between newlineselected seven data sets and cotton yield. The Root Mean Square Error (RMSE) a