[data mining] final review Chapter 1

Chapter one The introduction

1、 Definition of data mining

Technical level ： Data mining is From a lot of 、 Not completely 、 Noisy 、 Vague 、 In random practical application data , Extraction is implicit in 、 What people don't know in advance 、 But potentially useful information The process of .

Business level ： Data mining is a new business information processing technology , Its main feature is the analysis of business data extract 、 transformation 、 Analysis and other modeling processes , Extract key data to assist business decision-making .

2、 The task of data mining

Prediction task ： According to the values of other attributes , Predict the value of a specific attribute , Such as classification 、 Return to 、 Outlier detection .

Describe task ： Look for potential connection patterns in summary data , Such as cluster analysis 、 Correlation analysis .

• classification (Classification) analysis
  Classification analysis is to analyze data , Make accurate description for each category or establish analysis model or mine classification rules , Then use this classification model to classify other records .
• clustering (Clustering) analysis
  Cluster analysis technology is to find out the similarities and differences in the data set , And aggregate objects with common characteristics in the same class . Clustering can help determine which combinations are more meaningful .
• relation (Association) analysis
  Correlation analysis , Discover the co-occurrence relationship between things . It is usually to discover frequently occurring schema knowledge from a given data set ( Also known as association rules ). Relevance analysis is widely used in marketing 、 Transaction analysis .
• outliers (Outlier) analysis
  Outlier analysis is the mining of data points that deviate from most of the data . For example, the automatic detection of commercial fraud , Network intrusion detection , Financial fraud detection, etc .

3、 The object of data mining

Including spatial database 、 Time series database 、 Stream data 、 Multimedia database 、 Text data and the world wide web

4、 The main steps of knowledge discovery

• Data cleaning (data clearing). Clear data noise 、 Obvious with the mining theme irrelevant The data of .
• Data integration (data integration). Related data from multiple data sources Combine together .
• Data conversion (data transformation). Convert data to Easy data mining Data form of .
• data mining (data mining). Use intelligent methods to mine data patterns or laws .
• Model assessment (pattern evaluation). According to the evaluation criteria, meaningful relevant knowledge is selected from the mining results .
• Knowledge means (knowledge presentation). Using visualization and knowledge representation techniques , Show users what they have
  Mining related knowledge .

5、 The background and application field of data mining

The background ：“ Data surplus ”、“ Information explosion ” And “ Lack of knowledge ” Make people drown in data , Difficult to make the right decisions !

• The application of data mining in business field ： Market analysis and management , Corporate analysis and risk management , Fraud detection , Automatic trend forecasting ,…
• The application of data mining in computer field ： Intrusion detection , Spam filtering , Internet Information / Use mining , Intelligent response system …
• Applications in other fields ： Applications in industrial manufacturing , Data mining of biological information or genes ,…