Machine Learning test

My solution

• NN_for_Datarock_test.ipynb contains all work.

  • data exploration
  • data preparation
  • model definition
  • model training
  • model testing and output of results

• best_val_loss.pth contains a reference trained model which achieved MAE of 1.32 on Test set

• test_pred.csv is the prediction outputs

Test 1

• Download the train.csv and test.csv datasets from the test_1 folder.
• Each dataset has 11 columns: X1, X2, X3, X4, X5, X6, X7, X8, X9, X10 and Y.
• Build a deep neural network model from scratch (i.e. not using a canned or built-in model) to predict the Y values from the X* values.
• Train the model on the train.csv dataset and report the mean absolute error (MAE) of the model (it would be a good signal if the MAE is less than 2) .
• Run the model to predict the Y values for the samples in the test.csv dataset and save the results in the test_pred.csv file.
• Please provide your git repository, the test_pred.csv file