Abstract
Machine learning techniques have provided opportunities for researchers to design new analytical methods in different areas of agricultural extension. The objective of this study is to estimate the decision-making pattern of farm women in animal husbandry practices. Different machine learning algorithms namely linear regression, support vector machines, k-nearest neighbors, KStar, decision table, M5Rules, random forest and random tree were used for prediction of decision making pattern of farm women in animal husbandry practices in Junagadh district of Gujarat. The feature selection algorithm suggested that the input variable namely age, education, occupation, milk production, type of family, social participation, mass media exposure, extension participation, cosmopoliteness, scientific orientation, risk orientation, economic motivation and innovative proneness have significant influence on decision making pattern of farm women in animal husbandry practices. Based on all the benchmarks used to measure the predictability of fitted algorithms employed in this study, it was discovered that a support vector machines (SMOReg) performed better by achieving the highest prediction accuracy of 87%, lowest MAE of 2.78 and RMSE of 3.97 as compared with other fitted machine learning algorithms. Thus, the SMOReg algorithm explained 87 per cent of total variation in decision making pattern of farm women in animal husbandry practices. It was further observed that, the actual decision making pattern and the predicted decision making pattern of farm women were close to each other and the residual ranged from -4.36. to 12.47.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have