Application of Adaptive Neuro-fuzzy Inference System (ANFIS) for Prediction of Nutrients (Nitrogen, Phosphorus, Potassium and Magnesium) level from Soil Sample in some States of Nigeria Vegetational Zone. Abdullahi Yusuf Egwoh, Emmanuel Amano Onibere, Philip Odion

Abstract: Lack of soil nutrients knowledge can cause serious yield challenges if remains undiagnosed, misdiagnosed and untreated. It is the most prevalent problem identified by Nigeria Agricultural Development Project (ADP) to be affecting majority of the productive adult farmers in Nigeria. This study identified and compare qualitative and quantitative decision methods. Soil laboratory were used for determine soil nutrients level. This research proposes a hybrid soft computing model driven, by complimenting the advantages of Artificial Neural Network (ANN) and Fuzzy Logic (FL) with the use of Adaptive Neuro Fuzzy Inference System (ANFIS) for predicting the soil nutrients level. The Matrix Laboratory (matlab) applications were used for the analysis. The results of the analysis were found to be within acceptable predefined limits as examined by ADP officials. Evaluation of the model using standard statistical methods proved that the model is effective in providing accurate soil nutrients level (R² of 94%). The Root Means Square Error (RMSE) from ANFIS simulation is very low (0.000162) which indicate the significant and effectiveness of the model for prediction.

Keywords: Adaptive Neuro Fuzzy Inference System (ANFIS), Soil formation, Triangular membership function and Rule Viewer.