For each ML/AI model we create, we need some metric to assess how "good" or how "bad" the model is (how well or not the model is likely to perform). There are multiple metrics. R2 is a [0-1] value measuring the proportion of the variance in the data explained by the independent variables. If R2 == 1, then the independent variables perfectly explain the dependent variable - eg with linear regression, given x you perfectly predict y - the data looks like a straight line. If R2 == 0, then x gives no help in predicting y - the data looks random. In this sense, we can say the higher R2 the "better" the regression model. You can look up "coefficient of determination" - eg https://en.wikipedia.org/wiki/Coefficient_of_determination