KNN - Iris Prediction

Supervised Learning

• Learning from labeled data
• Introduction to scikit learn
• Works with numpy and scipy

1. Split the data set into two new datasets: training set (e.g 70% of the data sampems) and testing set oif the data
2. pretent that we do not know the label of the testeing set
3. train the model only on training set, and the predict on the testing set
4. after prediction, we can compare the predicted labels for the testing set with the actual labels of it to evaluate the accruacy of our prediction

Important things about sklearn

• only accepts numerical features. Thus, we have to convert non-numerical features into numerical values
• We use labelEncoding to convert labels an/or OneHotCoding to convert features

Used Knn

• Pros: No training phase—just store the data.
• cons: Slow for large datasets,