Data mining Getting Started algorithm collation

Recently is going to learn some knowledge of data mining, began to read some related blog, but too fragmented, has not a more systematic understanding of this. Weekend in the library wandering, accidentally saw "big talk data Mining" a book, found that the more organized, and quite suitable for the introduction, so read two chapters, make this note. This article simply introduces some algorithms for data mining, and does not involve the implementation of specific algorithms.
Here is the sorting algorithm classification diagram:

1. Association

Association refers to the dependency between an event and another event. The relationship between a classic diaper and beer.
The Apriori algorithm is the most classical association rule algorithm, the basic idea is: first from the event focused on finding all the frequent subsets. The higher-confidence rules are then found in these subsets.

2. Clustering

Clustering is the classification of objects into several classes, in the same class objects have a higher similarity, not the same kind of objects in a large difference.
The distance between objects can be divided into two types of algorithms: the Partitioning Method (partitioning methods) and the hierarchical Clustering Method (hierarchial). The former idea is: iterative calculation of the point to the center of the class distance, the data points within the class is better, the distance between the class as far as possible. The latter idea is: the object is layered into clusters, forming a cluster-node tree, the bottom-up is the aggregation of hierarchical clustering, top-down is the division of hierarchical clustering.

3. Forecast

Prediction is the process of predicting the future data by studying the correlation between the input and output of things, and then using the model to predict the forecast model. As a simple model of machine learning: input data for (X1,Y1), (x2,y2) ... (Xn,yn), by inputting the data, we can get β in model y=f (x,β), so we can determine the Y value by X for new data.
The specific prediction algorithm has many, the respective effect and realizes oneself Baidu Bar.

4. Sequence and Time series

A sequence is an object that is arranged in a column, that is, an ordered object. The time series is a sequence of time-separated sequences. Because the sequence is closely related to the association, the value in front of the sequence is likely to affect the value behind the sequence, which is similar to regression analysis for time series.

Listed here are just a few of the more representative introductory algorithms, and there are many algorithms or variants under each category, such as those commonly used in predictions and the proximity and naive Bayes, which are not given here. In addition, this article is mainly to give the concept of these algorithms, after reading may be unintelligible, so if you really want to understand, you need to check the corresponding algorithm related data.

Data mining Getting Started algorithm collation