How can I make my model more efficient?

Hello, I have this model: Google Colab

but I am not sure how reliable it is, so I am looking for help to tell me if I am doing it right or if there is something else I can modify, I appreciate your comments to help me.


Welcome to the Tensorflow Forum,

I think the way it’s done seems right, but I can’t say for sure if your results are good because I don’t have access to the real data. It’s important to evaluate how well the model works based on domain knowledge and the specific requirements of your problem.

To enhance the performance of the regression model, you can experiment with different architectures and hyper-parameters this includes adjusting the number of hidden layers, the number of neurons in each layer or trying different activation functions.

Thank you!

I think he means the accuracy, which isn’t good:

MSE: 12.757420224331531
R²: -0.2208057630939264

And seems to obtain constant line for the most part:

Predicción: 7.252624 Valor real: [13]
Predicción: 7.8349566 Valor real: [3]
Predicción: 7.2078295 Valor real: [7]
Predicción: 7.969341 Valor real: [5]
Predicción: 7.1182394 Valor real: [8]
Predicción: 7.700572 Valor real: [3]

One thing I’d do is to get rid of almost any extra package (scikit and pandas) apart from tensorflow, and use Splits and slicing  |  TensorFlow Datasets , I cant be sure it will work, but seems a leaner and clean approach.

I’d recommend you to add an open google drive link to the data, so that we can see where the problem lays exactly. It may be the way you normalise, or any thing else.


Adding to this, make sure that your training data covers entire distribution with all possible cases.

Thank you!