Dealing with negative values



I am trying to predict the electricity usage of heating and cooling appliances in a household based on internal and external temperatures. my data set includes negative values. how i can deal with the negative values?? keeping them as it s and building a model is good practice/. kindly help me in prediction.


Find the lowest negative value(or highest value with Negative sign) and add it as a constant for entire data set. This way all data is above 0.

Post prediction, do a reverse application of constant. This is one way and probably the best way.