The Prediction Report has a Technical Specifications tab which gives you detailed insights regarding the algorithm, columns used, model performance and accuracy metrics.
The Tech Specs contain the following Information:
- Details of the Identifier and Prediction Column.
Identifier Column is used to identify an entity/row. For example, if the dataset consists of different customers, the identifier column would be Customer ID. Prediction Column is basically the target variable that you want to predict.
- List of columns which were used for the analysis.
- Columns which are irrelevant and are discarded from the analysis and predictions.
- The algorithm used for prediction. Obviously AI tries out several state-of-the-art Machine Learning algorithms and chooses the best one for you.
- Performance of the model. This tells us whether the model is Excellent, Good, Moderate, or Bad.
- Training, validation, and testing accuracy.
- Percentage of the Data used for Training and testing. Training/Testing split 80/20, meaning 80% of the total data was used for training the model and 20% was used to test the performance of the model.
- Time taken to train the model.
- Loss Function used.
- Total Rows and columns in the dataset and the number of rows and the columns that were dropped from the dataset.