Introduction to ML Classification Algorithm
The process of guessing a category or a class from a given set of observations is known as Classification. The output can be categorized into "Yes" or "No" or "Red" or "Black." It makes a conclusion based on observable values. It is a kind of supervised learning, so the target is always given along with the dataset. It has a wide variety of applications ranging from medicine to marketing.
It basically approximates the mapping function (f) from the input variables to the output variables. One of the best examples would be the spam mail detection, as the outcome will be either categorized into the spam or the non-spam.
For this, we will require a classifier and will need to train it. In the case of spam mails detection, we will need to train the classifier with spam and no-spam emails, which will act as the training data.
Different types of learners in Classification
There are following types of learners in Classification:
Lazy learners: These are those learners who wait for a while for the test data to get available after the train data is stored. In this case, Classification is done after getting the test data. It takes much more time to make a prediction and spends very less time on training. Its examples are K-nearest neighbor and case-based reasoning.
Eager learners: These are the opposite of lazy learners as they do wait for the testing data to get appear after storing the train data. In fact, they spend more time training the data and less time on prediction. It includes Naïve Bayes, Artificial Neural Network (ANN), and Decision trees.
Types of ML Classification Algorithms
- Logistic Regression
- Support Vector Machine
- Decision Tree
- Naïve Bayes
- Random Forest
All these topics will be further discussed in detail chapters wise.
Some of the applications of classification algorithms are given below:
- Speech Recognition
- Handwriting Recognition
- Spam Email Classification
- Biometric Identification
- Emotion analysis
- Recognition of Cancer tumor cells, etc.