Akaike's Information criterion is a way to choose the best statistical model for a particular situation. According to the University of Georgia’s Fish & Wildlife Research Unit, the general Akaike’s Information Criteria (AIC) is calculated as AIC = -2_ln(likelihood) + 2_K. Once the AIC has been calculated for each model, further calculations are done to compare each model. These calculations involve calculating the differences between each AIC and the lowest AIC, and compiling this information in a table.

Calculate the number of model parameters. For example, the regression equation Growth = 9 + 2_age + 2_food + error has four parameters, while Growth = 2_age + 2_food + error has three parameters.

Multiply Step 1 by 2. Set this number aside for a moment.

Find the natural log of the likelihood.

Multiply Step 3 by -2.

Add Step 2 to Step 4.