Supervised learning is a very important machine learning algorithms. Compared with unsupervised learning, supervised learning have clear goals.
Classification and Regression two main tasks are supervised learning, unsupervised learning algorithm common mainly in the following nine kinds:
| 1 | Naive Bayes | classification |
| 2 | Decision Tree | classification |
| 3 | Support Vector Machines | classification |
| 4 | Logistic regression | classification |
| 5 | Linear Regression | return |
| 6 | Regression Trees | return |
| 7 | Nearby K | Category + return |
| 8 | Adaboost | Category + return |
| 9 | Neural Networks | Category + return |
1. Naive Bayes (Naive Bayes Classifier)
Naive Bayes algorithm (NBC) is one of the most widely used classification algorithm. NBC assumed between the attribute data sets are independent of each other, the classification used in the text.
2. The decision tree (Decision Tree)
Decision Tree algorithm uses a tree structure, the layers of reasoning to achieve the final classification.
Usually consists of a decision tree root, internal nodes, leaf nodes three elements, ID3, C4.5, CART decision tree is commonly used in three typical algorithms.
3. Support Vector Machine (SVM)
SVM classification problem into the problem of finding the plane's classification, and classification plane by maximizing the distance from the boundary points classification to achieve classification.
SVM can solve the problem of high-dimensional, machine learning can solve small sample size problem.
4. Logic Regression (Logistic Regression)
Logistic regression is a process dependent variable regression categorical variables, commonly two or binomial classification problems, may also handle multiple classification problems, it is actually part of a classification method, to represent something happened possibilities.
Simple logistic regression, classification when computing the amount is very small, very fast, low storage resource, mainly used in industrial problems.
5. Linear Regression (Linear Regression)
Linear regression is one of the most common processing tasks regression algorithm. In the form of the algorithm it is very simple, it is desirable to use a super-plane fit the data set (when only two variables is a straight line).
Linear regression modeling fast, it does not require complex calculations, in the case where a large amount of data is still fast speed, and can be given to understand and interpret a coefficient for each variable.
Logistic regression and linear regression are the following differences:
6. Regression Trees (Regression Tree)
Regression Trees, by definition, do regression model with the tree, every leaf is a predicted output value.
回归树通过将数据集重复分割为不同的分支而实现分层学习，分割的标准是最大化每一次分离的信息增益。这种分支结构让回归树很自然地学习到非线性关系。
7. K邻近(K-Nearest Neighbor)
K邻近算法是最简单的机器学习算法。
该方法的思路是：在特征空间中，如果一个样本附近的K个最近(即特征空间中最邻近)样本的大多数属于某一个类别，则该样本也属于这个类别。
K邻近算法理论相对成熟，思想简单，既可以用来做分类也可以用来做回归。
8. AdaBoost
AdaBoost目的就是从训练数据中学习一系列的弱分类器或基本分类器，然后将这些弱分类器组合成一个强分类器。
AdaBoost有一个很突出的特点就是精度很高。
9. 神经网络
神经网络从信息处理角度对人脑神经元网络进行抽象，建立某种简单模型，按不同的连接方式组成不同的网络。
在人工智能领域，神经网络通常指人工神经网络，即ANNs。