I am creating a fairly basic (3 independent variables) multivariate regression in TensorFlow.js following their simple docs found here: https://js.tensorflow.org/api/latest/#train.sgd
I have been relatively successful in creating predictions with outputs such as:
a: 0.34, b: 1.28, c: 0.72
. . .
i: 114, pred: -1.35, actual: 0.03, diff: 1.3800000000000001
i: 115, pred: 1.43, actual: 0.73, diff: -0.7
i: 116, pred: 0.18, actual: 0.01, diff: -0.16999999999999998
i: 117, pred: 0.41, actual: 0.41, diff: 0
i: 118, pred: -2.58, actual: 0.59, diff: 3.17
i: 119, pred: -3.12, actual: 0.61, diff: 3.73
i: 120, pred: 1.58, actual: 1.25, diff: -0.33000000000000007
i: 121, pred: 4.72, actual: 3.19, diff: -1.5299999999999998
i: 122, pred: 8.93, actual: 7.61, diff: -1.3199999999999994
However, at a glance I cannot tell the significance of the variables or how good of a fit the model is as a whole.
Are there any built in methods to utilize before I go digging in to manually calculate standard deviations and variances?