Multi-regression

Now that you have learned how to create a basic regression model with TensorFlow, let's try to run it on example datasets from different domains. The dataset that we generated as an example dataset is univariate, namely, the target was dependent only on one feature.

Note

Most of the datasets, in reality, are multivariate. To emphasize a little more, the target depends on multiple variables or features, thus the regression model is called multi-regression or multidimensional regression.

We first start with the most popular Boston dataset. This dataset contains 13 attributes of 506 houses in Boston such as the average number of rooms per dwelling, nitric oxide concentration, weighted distances to five Boston employment centers, and so on. The target is the median value of owner-occupied homes. Let's dive into exploring a regression model for this dataset.

Load the dataset from the sklearn library and look at its description:

boston=skds.load_boston()
print(boston.DESCR)
X=boston...