Application of Artificial Neural Network modeling for optimization and prediction of essential oil yield in turmeric (Curcutna longa L.)

The essential oil obtained from rhizome of turmeric (Curcuma longa L.) is highly valued worldwide for its medicinal and cosmetic uses. Lack of requisite high oil containing genotypes and existing variation in the quality and quantity of essential oil with plant habitat and agro-climatic regions pose problem in commercialization of essential oil. Thus the present work was carried out for optimization and prediction of essential oil yield of turmeric at different agro climatic regions. An artificial neural network (ANN) based prediction model was developed by using the data of essential oil of 131 turmeric germplasms collected from 8 agro-climatic regions of Odisha and analysis of their soil and environmental factors. Each sample with 11 parameters was used for training and testing the ANN model. The results showed that multilayer-feed-forward neural networks with 12 nodes (MLFN-12) was the most suitable and reasonable model to use with R-2 value of 0.88. This study indicates that ANN based prediction model is a suitable way of predicting oil yield at a new site and to optimize the yield of turmeric oil at a particular site by changing the changeable parameters of the prediction model and thus is of enough commercial significance.