SUPERVISED MACHINE LEARNING ALGORITHM

Supervised Learning is one of the successful types of Machine Learning. Remember that supervised learning is predicting the outcome from given input where we have examples of input-output pairs.

The Goal is to “predict the outcome for unseen data”.

One important thing about supervised Machine Learning is it requires human effort to build a training set ( input-output pairs), but later automates and often speeds up the task.

There are two types of Supervised Machine Learning Algorithm

• Classification
•  Regression

Classification:

The goal of Classification is to predict a class label (which can be termed as outcome sometimes).

Classification can sometimes be distinguishing between exactly two classes which are called binary classification and classification between more than two classes are called multiclass classification.

Binary classification can be referred to as answering a YES/NO question. For example, Is this email spam? The answer is either yes or no.

Example of Multiclass classification can be predicting what language a website is in from the text on the website.

Regression:

The goal of Regression is to predict a continuous number or a floating-point number.

For example, predicting the yield of corn farm given attributes such as weather, previous yields, and number of members working on the farm. 

Another example could be predicting the person’s annual income given their education, age, and where they live. And the most common example used for representing regression is predicting the house price given attributes number of beds, size, number of bathrooms.

Image Source: Google

An easy way to distinguish between classification and regression task is to ask whether “Is there some kind of continuity in the output”. Well if there is continuity between possible outcomes, then the problem is a regression problem. Let’s look at the example discussed above, there is continuity in the output where we predict the annual income, it does not make a tangible difference whether a person makes $90,000 or $90,001 a year. By contrast, for recognizing the language on the website which is other than a classification problem), the output matters most. There will be no continuity between languages and there is no such language between English and French.

Supervised Machine Learning Algorithms:

Nearest Neighbors – For small datasets, good as a baseline and easy to explain.

Linear Models – Go-to as a first algorithm to try and good for very large datasets and also good for very-high dimensional data.

Naïve Bayes – Used only for Classification. Even faster than linear models and good for very large datasets and high dimensional data. Often less accurate than linear models.

Decision Trees – Very fast, don’t need scaling of data, can be visualized and easily explained.

Random Forests – Nearly always perform better than a single decision tree, very robust and powerful. Do not need scaling of data and not good for very high dimensional sparse data.

Gradient boosted decision trees – Often slightly more accurate than random forests. Slower to train but faster to predict than random forests and smaller in memory. Need more parameter tuning than random forests.

Support vector machines – Powerful for medium-sized datasets of features with similar meaning. Require scaling of data, sensitive to parameters.

Neural networks – Can build very complex models, particularly for large datasets. Sensitive to scaling of the data and to the choice of parameters. Large models need a long time to train.

We will be going to discuss each and everything in the next posts.

I hope you enjoy reading this and I hope you understand what is a supervised machine learning algorithm and its types.