How do I predict a category of data for a future date ? Example: what will the Sales figure for region (or region wise) for a particular date in the future based on the sales person past data for a particular product and product category? which model will be the best for these kind of problem ?
Linear regression works well for predictive analytics if you data is linear and doesn't contain outlier i.e In places where you want to predict future trend depending on the the past/current trend, Linear regression does a fine job. Linear equation is
w0 + (wT*X) where X is your input,
w is the slope of your regression line and
w0 is the y-intercept of your regression line. The objective function of linear regression is an optimization problem where we try to minimizes
Check out: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html