2. Regression is a widely used statistical analysis method. By specifying dependent variables and independent variables to determine the causal relationship between variables, a regression model is established, and the parameters of the model are solved according to the measured data, so as to evaluate whether the regression model can fit the measured data well. If it can be well fitted, further prediction can be made according to the independent variables.
3. Clustering is a classification method, which divides data into some aggregation classes according to their inherent attributes. The elements in each aggregation class have the same characteristics as much as possible, and the characteristics of different aggregation classes are as different as possible. Unlike classification analysis, the category of classification is unknown. Therefore, cluster analysis is also called unsupervised or unsupervised learning.
Data clustering is a static data analysis technology, which is widely used in machine learning, data mining, pattern recognition, image analysis and biological information.
4. Similarity matching Similarity matching is to calculate the similarity of two data through a certain method, and similarity is usually measured by percentage.
Similarity matching algorithm is used in many different computing scenarios, such as data cleaning, user input error correction, recommendation statistics, plagiarism detection system, automatic scoring system, web search and DNA sequence matching.
5. Frequent itemsets Frequent itemsets refer to frequent itemsets in cases, such as beer and diapers. Apriori algorithm is a frequent itemset algorithm for mining association rules. Its core idea is to mine frequent itemsets through two stages: candidate set generation and closed graph detection, which has been widely used in business, network security and other fields.
6. Statistical description Statistical description is the basic processing work of data analysis, which displays the information of data feedback with certain statistical indicators and indicator systems according to the characteristics of data. The main methods include: calculation of average index and variation index, diagram of data distribution mode, etc.
7. Link prediction Link prediction is a method to predict the relationship between data. Link prediction can be divided into prediction based on node attributes and prediction based on network structure. Link prediction based on node attributes includes analyzing the relationship between nodes' attributes, attributes between nodes and other information, and obtaining the hidden relationship between nodes with the help of node information knowledge set and node similarity.
Compared with link prediction based on node attributes, network structure data is easier to obtain.
A main viewpoint in the field of complex network shows that the characteristics of individuals in the network are not as important as the relationship between individuals.
Therefore, link prediction based on network structure has attracted more and more attention.
8. Data compression Data compression refers to a technical method that reduces the amount of data to reduce storage space and improve its transmission, storage and processing efficiency without losing useful information, or reorganizes the data according to a certain algorithm to reduce data redundancy and storage space.
Data compression is divided into lossy compression and lossless compression.
9. Causal analysis Causal analysis is a method of forecasting by using the causal relationship of things. Causality analysis is used to forecast the market, mainly through regression analysis. In addition, the calculation of economic model and input-output analysis are also commonly used.