- Model – A file that has been trained to recognize certain types of patterns.
- Algorithm – a set of clearly defined instructions a computer uses to solve a problem such as learning and prediction
- Hyperparameters – the variables that are used to configure an algorithm for learning and prediction
- Model Tuning – changing hyperparameters to improve the performance of a model
Types of Models
- Classification – when the target/output is discrete or categorical e.g Jacket or no Jacket, good or bad etc.
- Regression – when the target is continuous e.g temperature – 15.8,19.6,12.0 or salaries 100K, 60K, 80K etc.
- Clustering – when the objects in the data has to be grouped to belong into similar distributions e.g. afternoons are always hot, etc.
Model vs. Algorithm
- An algorithm is the set of instructions a computer follows to solve a problem.
- A model is the set of assumptions formed when an algorithm that takes some values as input and produces some values as output.
Building a Model
- Use different algorithms to train a model on the dataset
- Test the model’s performance using the testing data
- Improvement performance of the model by tuning the hyperparameters
Now learn how to evaluate model performance.