A Machine Learning Approach for Stochastic Pattern Analysis for the Measurement of Time-Series Datasets

Abstract

Time series data analysis is used in many domains of applications to perform efficient prediction, anomaly detection and forecasting using various techniques. Recently, there is a surge in demand for comprehensive data mining techniques that can be applied on agricultural prediction. Data mining techniques can be used effectively for predicting the agricultural growth using stochastic model sensing concept which is used to perform better analysis for predicting the dataset. The proposed work introduces a stochastic model that can be applied for agricultural data to predict the growth of any crops using primary time series datasets The framework has assisted in taking correct decisions to classify the dataset based on the threshold value of micro and macro nutrient obtained from National Food Security Mission (NFSM). This paper focus on analyzing the stochastic pattern that assist the proportion of micronutrient elements and also predicts the feature that affect the growth of agriculture by increase or decrease the level of micronutrient elements. Thus, the Expected Growth of Agriculture (EGA) has increased or decreased based on the strength of soil nutrients. Moreover, it will recommend chemical fertilizers and nutrients range to improvise the agriculture growth for each crop. The processed soil sample dataset based on the threshold value of micro nutrient level is obtained from National Food Security Mission (NFSM). There are several numerical illustrations that are performed for micronutrient data prediction as well as analyze the increase or decrease of growth level in agriculture based on micronutrient levels like Fe, Mn, Zn and Cu using Stochastic Weibull Distribution (SWD) model.