# Classic algorithm vs ML algorithm

## Classical algorithms

Classical algorithms are step-by-step instructions such that given specific input one can trace and determine exactly the output (ignore for now randomized algorithms). You can think as if the algorithm is a rule-based one.

For example, assume we need to sort N numbers. We can write instructions for the computer to sort such numbers. We know exactly what will be the output for a set of N numbers.

## Machine learning algorithms

Actually, the machine learning field comes to help. It tells you no more explicit instructions for writing. The ML algorithm will learn from data and encode such spam detector implicitly.

## Differences between Classic algorithms and ML algorithms

• You can think about many problems where writing classical algorithm is very tough. However, if you can collect enough data, a machine learning algorithm can learn to find the answer implicitly.

• Machine learning algorithm can learn from data while classical algorithms specify the exact rules to find the overall answer.

• An ML algorithm is an algorithm intended to be used in ML.

• A classic algorithm uses code and data to predict the correct answer for a question. On the other hand, an ML algorithm focuses on changing that code and/or data in a mostly automated way in order to make better predictions.

• The classic algorithm produces an output on the basis of steps described in the algorithm. In contrast, the ML algorithm predicts an output on the basis of learning through the input provided to it. This learning through input is called the Training process.

• Machine learning algorithms are a class of computer algorithms related to Newtonian estimation for determining the square root of a number and linear regression in statistics.

• In Classic algorithms, it typically takes a large number of examples to determine how far the equation is from the desired equation. This is why "big data" is a big deal right now.

• Machine learning algorithms typically solve for the coefficients of hundreds or thousands or even tens of thousands of dimensions.

• A classic algorithm doesn't give a solution after reaching an optimum solution for a problem. On the other hand, Machine learning algorithms give the new thoughts new ideas, new solutions after reaching optimum value for a problem.

• Machine learning algorithms have achieved accuracies that are far beyond that of classical algorithms.

• Machine learning algorithms scale much better with more data than classical algorithms.

Lastly say, both styles are important to our world. You can share your experiences with us. Thank you!