Overview / Usage

We tried Xtreme gradient boosting, support vector machine, relevance vector machine, variance optimized bagging,multiple linear regression and noticed the accuracy did not cross 90%, but random forest regression gave an accuracy of 98%.

We concatted both user data values and then inputed the missing values ( NaN ) as the column values were not the same and to prevent loss of values. Then we scale it down and on applying the random forest resgression we get the predicted value and then get an accuracy of 98% and test accuracy is 93%. Now to predict more data for the rest of march , we first find the mean and get the standard deviation with which we randomly generate input values for the next 20 days and predict the soil moisture , humidity and temperature of the soil

Repository

https://github.com/jaishana25/Soil-Moisture-Prediction