The coefficient of determination, a concept of statistics Also called R squared (or R²), it represents the proportion of total variance of the variable that explains the regression. In other words, this term tries to explain the goodness of the fit of a model to the variable it intends to analyze.
This coefficient can offer values between 0 and 1. If the result is 1 or close to 1, we can indicate that the model and the variable that we want to explain are highly adjusted. Conversely, if you get closer to 0, the model will fit less (in addition to being less reliable).
How is the coefficient of determination interpreted?
For the interpretation of said coefficient of determination, the variables that are being taken as reference in the model must be taken into account. Normally, the estimate that is usually taken as a reference is increasing or decreasing at a slow or fast level, depending on what is being measured.
In any case, it is intended to measure the distance that exists between the points of the graph (which will determine the different observations found on a certain experiment) in relation to that estimate made (which, as we say, can be positive - increasing - or negative - decreasing-).
When interpreting the result, as we have already indicated, it usually oscillates between 0 and 1. If, for example, we have a coefficient of 0 we can say that the model has estimates (or points that in the graph) that fit well to the real variable or that has been taken as a reference (the estimate). Furthermore, although it would not be entirely correct to say so, we can indicate that the model explains 9% of the real variable that has been taken as a reference.
As the main drawback, we must take into consideration that this model does not take into account that explanatory variables are included that have little to do with the model or the variable to be explained.