Algorithm

Simply put, an algorithm is a general set of instructions created to solve a specific type of problem. It is considered to be accurate when it produces the correct output for each problem instance. Let’s specify some of the used terms to make this definition clearer:

• In computer science, every problem has a set of instances (which can be described by several parameters) and a question about these instances
• An instance of a problem is obtained by specifying precise values for all parameters. The values of the parameters that make up an instance can also be called input data or simply entered
• The response obtained at the end of the execution is called algorithm output
• Any execution of the algorithm must end (after a number of steps) and give a correct result (the answer to the question)

Why do data scientists use algorithms?

The typical data scientist has only one idea in mind while using algorithms - discovering links in his data (also known as patterns). In the context of using data science methods, we always assume that there are hidden links in the data and that our algorithms will help us find them. However, you need to keep in mind that this is a strong assumption because there is usually no link in data by default. This may seem surprising, because we tend to see it everywhere!

Examples of algorithms and utilities in data science:

Support Vector Machine (Regression or classification); Gradient Boosting (Regression or classification); Random Forest ( Regression or classification); Polynomial Regression ( Regression)

Concrete examples:

Logistic Regression : Spam Detection (Predicting if an email is Spam or not ); Decision Tree: Churn prediction; Random Forest: Classification of customer tastes

Additional Resources:

• A Quick Guide to Classification Algorithms and How to Chose the Right One