Regression model
The forecast applied to the data.
Regularization model
By introducing penalties, the model can be prevented from over-fitting and the generalization of the model can be improved.
Decision tree model
It can be used for prediction and classification.
Integrated model
Integrate multiple weak models together to greatly improve the generalization and accuracy of the model. And naturally avoid overfitting of the model.
Typical: Random Forest Random Forest
Clustering Algorithms
Judge which samples are of the same class by distance measurement.
Typical: K-Means clustering method
Classification model Instance-based Algorithms
Given a proxy test sample, determine which category the proxy test sample belongs to.
Typical: KNN method
Support Vector Machines Model Support Vector Machines
Used for classification of city dimension data, such as image classification, face recognition, etc.
Graphical models
Applied to path problem
Association Rule Learning Algorithms
Take the case of Wal-Mart, beer and diapers.
Bayesian Algorithms
The bottom layer is realized by Bayes' theorem. Applied to the filtering of spam and the inference of some problems.
Dimensionality Reduction Algorithms
Used to eliminate unimportant dimensions, thereby reducing computational cost
Recommender system model
Artificial Neural Network
Deep learning model
After the model is established, the coefficients need to be solved, which can be solved by some algorithms, such as:
least square method,
stochastic gradient descent method,
batch gradient descent method,
Newton method
ID3 algorithm
C4.5
by the Yees algorithm