Is linear regression a better model than a regression neural network?

In the last few posts I have been working on Jax regression neural networks in an attempt to solve machine learning projects. The link to the most recent blog post I have written, which uses a regression neural network to predict on the Boston house price dataset can be found here:- https://tracyrenee61.medium.com/use-jaxs-regression-neural-network-to-predict-on-boston-house-prices-fef89b4f9321

In this blog post I have compared the linear regression model with the regression neural network to see which one works the best to predict house prices. I have used the Boston house piece dataset, which can be found in my previous blog post, cited above. This dataset is no longer on Python’s machine learning library, sklearn, so I had to use the csv file on Jason Brownlee’s Github account.

Below is a diagram of a regression neural network, which has the same architecture as a classification neural network, except it is lacking an activation function on the output node:-

When I trained and tested the regression neural network using the Boston house price dataset, I had achieved a root mean squared error (RMSE) of 9.98:-