Table of contents
1. What is Machine Learning
from a melon
Gain experience
use experience
new decision
2. Classification and regression problems
characteristic | classification problem | regression problem |
---|---|---|
Output type | discrete data | continuous data |
Purpose | Finding the decision boundary | find the best fit |
binary classification
linear regression
polynomial regression
clustering problem
Note: The label is not given during the training process in the clustering problem.
supervised learning
Supervised learning is to "learn" a function from a given training data set. When new data arrives, the result can be predicted based on this function. The training set requirements of supervised learning include input and output, that is, features and targets. The objects in the training set are annotated by humans in advance.
Main usage: classification problems and regression problems
Common algorithms: decision trees and random forests, logistic regression,
neural networks, naive Bayes, logistic regression, etc.
unsupervised learning
In the process of unsupervised learning, only the attributes of things are provided, but the labels of things are not provided, allowing learners to summarize by themselves. So unsupervised learning, also called inductive learning, refers to the process of dividing a data set into multiple clusters (or groups) composed of similar objects. In order to achieve the result that the characteristics of things in the same group are very close, and the characteristics of things in different groups are far apart.
Main use: clustering problems
Common algorithms: k-means, Apriori, FP-Growth
semi-supervised learning
reinforcement learning
Reinforcement learning is to learn an optimal policy that allows the agent to take actions based on the current state in a specific environment to obtain the maximum reward.