Unsupervised discretization method for continuous attribute data based on information entropy
A technology of attribute data and information entropy, applied in the field of unsupervised discretization of continuous attribute data based on information entropy, can solve the problems of high calculation cost, lack of theoretical basis, lack of data adaptability, etc., and achieve the effect of high calculation efficiency
Inactive Publication Date: 2021-02-12
NORTHWEST NORMAL UNIVERSITY
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0006] Among the above methods, the method of subjectively specifying the number of discrete values by the user lacks adaptability to the original data; the method of assuming conditions lacks theoretical basis; the discretization process depends on other attributes through the heuristic method; the method of using information entropy is not based on The information entropy of continuous attributes determines the number of discrete values, and the calculation cost is relatively high
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0029] The present invention will be described in detail below in conjunction with the accompanying drawings.
[0030] First, a brief description of the terms and discretization process involved in the present invention is given to facilitate the understanding of the method of the present invention.
[0031] Discrete granularity: For a given data set, the number of different values of any attribute is called the discrete granularity of the attribute. The discrete granularity of discrete attributes is denoted as |c|, and the discrete granularity of continuous attributes is denoted as |n|.
[0032] Information entropy is a theory for measuring information uncertainty, which is defined as the probability of occurrence of discrete random events. Similarly, each different value of a continuous attribute can be understood as a discrete random event, and the discrete granularity of this attribute is equal to the number of discrete random events. By calculating the information ent...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention relates to the technical field of discretization of continuous attributes of big data, in particular to an unsupervised discretization method for continuous attribute data based on information entropy. The method includes the steps as follows: the first step of traversing all value records of any continuous attribute, counting discrete granularity |nj| of the attribute and the probability qji of each different value, and recording a maximum njmax and minimum njmin; the second step of obtaining a calculating formula of the value chaos degree of any continuous attribute nj according to a calculating formula of the information entropy, and calculating the value chaos degree of the attribute according to the formula; the third step of rounding down the value chaos degree to obtain the number of break points; the fourth step of adopting an equivalent width interval method to calculate the width of each divided interval, and determining the position of each break point; and thefifth step of discretizing the continuous attribute nj. The novel method for determining the number of break points is more suitable for the original data, and the discretization of each attribute does not affect each other and does not depend on other attributes, and the calculation efficiency is higher.
Description
technical field [0001] The invention relates to the technical field of discretization of continuous attributes of big data, in particular to an unsupervised discretization method of continuous attribute data based on information entropy. Background technique [0002] Discretization of continuous attributes is the process of dividing the value range of continuous attributes into several intervals, each interval corresponding to a unique discrete value, and transforming the original value into a discrete value. Researchers at home and abroad have proposed a lot of methods in the discretization of continuous attributes (numerical attributes), and there are many classification methods from different perspectives, namely top-down and bottom-up, supervised and and partial, static and dynamic, and single-attribute and multi-attribute, etc. The essence of discretization of continuous attributes is to determine the number of discrete values (intervals, intervals) and the location ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06F17/18
CPCG06F17/18
Inventor 马生俊陈旺虎郭宏乐乔保民李新田
Owner NORTHWEST NORMAL UNIVERSITY