What is Supervised Machine Learning?
As the name suggests Supervised means there is some supervision in the learning methodology. It is like learning from the teacher or having supervision from the master. The machine learning technique in which well-labeled data is used to classify or predict the correct output by learning from the previous similar attribute in the data. Well labeled means data has an input-output sequence. For this at first, some example training dataset is fetched to create a model then actual data is given to predict with that model. In weather forecasting scientist could predict it would rain or not by analyzing the attributes like temperature, pressure, wind, and humidity of similar condition before. In simple words, supervised machine learning is to classify or predict the output based on the example input-output pairs.
In simple words, supervised machine learning is to classify or predict the output based on the example input-output pairs.
Supervised learning can be further grouped into classification problem and regression problem.
Classification Problem
A classification problem is when the output variable takes class labels (category) or discrete value.
Some of the example problems that are solved by classification method are:
- From the given picture, classify if the person in the image is male or female.
- From the given drink, classify if it is Pepsi or Coca-Cola.
- Classify the given transaction in the bank is Fraud or not.
- Classify an email received is spam or not.
Regression Problem
A regression problem is when the output variable takes real value (continuous value) which is ‘dollar value $35.5′,’weight 50.5kg’ or ‘height: 165 cm’.
Some of the example problem solved by Regression method is
- Predict the price of the stock of AAPL (Apple Inc.) based on the previous stock attributes.
- Predict the price of a house in Kathmandu with 3 bedrooms, 1 kitchen and 1 living room in 200 sq. meter area.
- Predict the height of the 16-year-old boy in Nepal.
Algorithms for Supervised Machine Learning
Classification Algorithms
- Decision Trees
- Logistic Regression
- Naive Bayes
- K Nearest Neighbors (Knn algorithm)
- Linear SVC (Support vector Classifier)
- Neural Networks (Multilayer perceptron).
Regression Algorithms
- Linear Regression
- Regression Trees(e.g. Random Forest)
- Support Vector Regression (SVR)
Supervised Machine Learning is found to be more accurate than unsupervised learning because of the labeled dataset. You can learn more about Unsupervised Machine Learning here.
If you have still stuck or want to learn more about Supervised Machine learning than go through this great video of Andrew Ng.