1. Pattern tracking

Pattern tracking is a basic technology of data mining. It is designed to identify and monitor trends or patterns in data to form intelligent inferences about business outcomes. For example, businesses can use it to identify trends in sales data. If it is discovered that a certain product sells better than others among certain groups of people, the company can then create similar products or services based on that, or even simply increase the inventory of the original product for those groups.

2. Data cleaning and preparation

As an important part of the data mining process, we must clean and format the original data for various subsequent analyses. Specifically, data cleaning and preparation work includes various elements such as data modeling, transformation, migration, integration and aggregation. This is a necessary step to understand the basic characteristics and properties of the data to determine its best use.

3. Classification

Classification-based data mining technology mainly involves analyzing the associated attributes between various types of data. Once the key characteristics of data types are identified, businesses can classify them. Businesses can use this to determine whether to protect or delete certain personally identifiable information.

4. Outlier detection

Outlier detection can be used to identify anomalies in the data set. After discovering outliers in the data, enterprises can prevent such events from occurring to successfully achieve business goals. For example, if the credit card system has a peak in usage and transactions during a specific period, the company can understand through analysis that it may be due to a "big promotion" and make pre-deployment of resources for such activities in the future. Prepare.

5. Correlation

Correlation is a data mining technology related to statistics. It aims to establish connections between certain data and other data, or data-driven events. It is similar to the concept of "co-occurrence" in machine learning, that is: the probability of an event based on data is identified by the existence of another event. For example, the user's behavior of purchasing a burger is often accompanied by the possibility of purchasing potato chips. There is a strong correlation between the two, but they are not absolutely associated.

6. Clustering

Clustering is an analysis technique that relies on visualization methods to understand data. The clustering mechanism uses graphics or colors to show the distribution of data under different categories of indicators. Through graphical cluster analysis, users can intuitively understand the trend of data development with business goals.