Machine learning algorithms can be generally divided into two categories (if regardless of reinforcement learning): supervised learning and unsupervised learning. Supervised learning algorithms can be used to solve classification problems (the answer for a data is categorical, 0, 1, 2, …) or regression problems (the answer for a data is numerical)
So my recommended three algorithms are corresponding to the categories/problems each other, including:
K-means cluster, an unsupervised learning algorithm which “aims to partition n observations into k clusters in which each observation belongs to the cluster with the nearest mean, serving as a prototype of the cluster”. When I was a freshman, one of my programming assignment for the C++ programming language course is to implement it
Linear regression, a supervised learning algorithm for regression problems. If you know some matrix/vector calculus, the derivation of the final solution is very straightforward. Even if you don’t know anything about it, given the final solution you could easily implement it because this algorithm has a closed-form, analytical solution which needs only one line of code if you use some scientific calculation libraries
K-Nearest Neighbors algorithm, (usually also called KNN for short), a supervised learning algorithm for classification problems. Given new data, you just need to calculate the distances (which can have multiple definitions as the requirements could be different) between it and all your known data, pick out the k known data that have the smallest distances, and select the majority label. This algorithm can also be used to solve regression problems after a slight modification
Machine Learning is a subset of Data Science and even Machine Learning is applied in every possible area that's why people are also eager to have Machine Learning online course
Back