Hooda, B.K.Bharti, Deepa2020-02-042020-02-042019http://krishikosh.egranth.ac.in/handle/1/5810142438In the present study, growth rates in area, production, productivity and export of onion in India and world have been studied during 1980-2016. Support Vector Machine (SVM) and Artificial Neural Network (ANN) models were also developed and compared for monthly arrivals and prices of onion in major Indian Markets for the period 1990 to 2016. Study revealed a significant increase in area, production, productivity and export of onion in India during the period 1980-2016 with compound growth rates of 5.13, 6.72, 1.47and 7.62 per cent per annum. While significant increase with compound growth rates 3.54, 4.21 and 0.18 per cent per annum was also observed in area, production and productivity at the world level for the period 1980 to 2016. SVM was found to be the better predictor for onion arrivals in Delhi, Kolkata and Bengaluru markets. While, ANN performed better for the Mumbai and Jaipur market. Also for onion prices of selected market, SVM models were better predictors in three markets prices, while ANN models were fitted on two markets. In Kolkata, Bengaluru and Mumbai markets prices, lower value of RMSE and MAPE were found using SVM model than ANN model. For the Jaipur and Delhi market, least value of RMSE and MAPE were found for ANN model. Hence SVM models outperformed ANN models.ennullThesis