Feature extraction method and device
A feature extraction and feature word technology, applied in the field of feature extraction methods and devices, can solve the problems of poor address text mining effect, and achieve the effect of improving the mining effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
example 1
[0096] Such as Figure 3A and Figure 3B Shown are the schematic diagrams of the process of generating feature word strings in the case of s<n.
[0097] In this example, set n=6, s=3, and the word in the current position is the i-th word from the start word of the address text (that is, the current position is i), then the gram generation process is as follows:
[0098] 1. Select n (n=6) words continuously from the word at the current position and put them into the buff. At this time, the words in the buff are put together to form the first word string (corresponding to the above step B1);
[0099] 2. Select s words (corresponding to the above-mentioned step B2) continuously from the first word (ie corresponding position i+n) after the first word string as the second word string;
[0100] 3, according to the difference of the initial jump word position in the first word string, can be divided into two kinds of situations again: in the first word string, the word quantity q o...
example 2
[0105] Such as Figure 4 As shown, it is a schematic diagram of another generation process of gram in the case of s≥n provided by Example 2.
[0106] In this example, set n=4, s=5, the word of current position is the i-th word (that is, the current position is i) from the start word of address text, then the generation process of gram comprises as follows:
[0107] 1. Select n (n=4) words continuously from the word at the current position and put them into the buff. At this time, the words in the buff are put together to form the first word string (corresponding to the above step B1);
[0108] 2. Select s words (corresponding to the above-mentioned step B2) continuously from the first word (ie corresponding position i+n) after the first word string as the second word string;
[0109] 3. In the case of s≥n, no matter where the initial jump word in the first word string is, the word quantity q of the nth word from the start jump word to the first word string in the first word s...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More