Project 1: Implementation of a Fully Connected Neural Network (FCNN)

Results

Model Build and Parameters

The FCNN has been created successfully and an image of the neural network architecture has been stored along side the model, parameters, features and scalers information. For reproducability all parameters are listed here:

  • Random State = 1234
  • Test Size (split percentage) = 0.3
  • Hidden Layers = 3
  • Neurons per Layer = 128
  • Epochs = 50
  • Batch Size = 32
  • Validation Split (split percentage) = 0.2

    • Model Evaluation

    Using these paremeters with the bike sharing dataset resulted in following model evaluation metrics:

    FCNN Model

  • Mean Squared Error (MSE: 1.90): A lower MSE indicates better model performance. In this case, the MSE suggests that the model's predictions are close to the actual values.
  • Mean Absolute Error (MAE: 1.03): An MAE of 1.03 indicates that, on average, the model's predictions deviate from the true values by approximately 1 unit
  • R² Score (0.9999): The R² score indicates the proportion of variance in the target variable explained by the model. An R² value of 0.9999 means the model explains nearly 100% of the variance, indicating an excellent fit.

    • Result Interpretation

    By comparing the expected results with the predicted values, we can evaluate the model's performance. In this case, the predictions closely align with the actual values, indicating strong accuracy. The small deviations observed are reflected in the evaluation metrics presented earlier. Overall they confirm that the model achieves an excellent fit to the data. More finetuning could further icnrease the model accuracy.


    Dataset Predicted Value (FCNN Result) Expected Value (Control Value) Difference (FCNN Deviation)
    1 94.5368423461914 93 1.5368
    2 68.8176498413086 67 1.8176
    3 35.34605407714844 35 0.3461
    4 36.78642272949219 37 -0.2136
    5 35.923160552978516 36 -0.0768
    6 32.77050018310547 34 -1.2295
    7 27.173233032226562 28 -0.8268
    8 38.892486572265625 39 -0.1075