A typical ML project follows a repeated cycle:
1. Data Representation
Decide how to present your data to the algorithm, such as pixels for images or numerical features for transactions.
This process is called feature engineering.
2. Choose an Algorithm
Select a suitable model (e.g. classifier or regressor) based on the type of problem.
3. Evaluation
Measure how well the model performs using metrics like accuracy.
This lets you compare models fairly.
4. Optimisation
Tune hyperparameters such as learning rate or model depth to improve results.
This process is repeated many times. Each cycle helps refine the features, improve model choices and increase accuracy.