Crop Suitability Prediction Model for Malaysian Crop Diversification

Crop diversity is one of the important perspectives to be observed in agriculture. The crop diversity is significant for production stability as well as nutrition security. However, the crop diversity in Malaysia is inadequate as the agricultural activities are devoted to oil crop plantation. Therefore, this research aims to discover the suitable new tropical crops to be cultivatable in Malaysia for crop diversification. FAOSTAT was selected as data source, while Decision Tree, Random Forest and Artificial Neural Network were chosen for predictive model training. The Random Forest is having the highest accuracy among the modelling techniques. Therefore, Random Forest models were chosen as crop suitability predictive models to discover the suitable crop with Malaysian environmental data as input. There are nine tropical crops subjected to investigation. The crops are dates, sorghum, yams, avocados, kola nuts, chickpeas, lentils, sisal and fonio. Dates, sorghum, yams, avocados and kola nuts were predicted to be not suitable to the Malaysian environment. Whereas chickpeas, lentils, sisal and fonio were predicted to be cultivatable with the Malaysian environment. Unlike other crop diversification research that done on other countries, which are soybean cultivation in Europe, red kidney bean in Iraqi and Cassava in Thailand. This research investigated the potential of nine tropical crops besides soybean, red kidney bean and cassava. At the same time, the crop suitability was predicted with respect to Malaysian environment. Future works are suggested to investigate the procedure of new crop cultivation and the tactic to release the crop to the market.