A Method for Solving Feature Extraction Tasks Using Image Regularization and Data Reconstruction
A technology of feature extraction and data reconstruction, applied in the fields of electronic digital data processing, digital data information retrieval, special data processing applications, etc., can solve the problem of not paying attention to the preservation of data structure information, etc.
Active Publication Date: 2021-12-10
ZHEJIANG UNIV
View PDF9 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] The purpose of the present invention is to solve the problems in the prior art, in order to overcome the problem that supervised feature extraction in the prior art requires a large amount of labeled data, and the existing unsupervised feature extraction only pays attention to the preservation of data characteristic information, but Without paying attention to the problem of preserving data structure information, the present invention provides a method for solving feature extraction tasks using image regularization and data reconstruction
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
[0059] The present invention constructs experimental data on the TDT2 corpus data set and the Reuters corpus data set for experimentation. The TDT2 corpus dataset used includes a total of 10,021 file data, and each file data belongs to a category; the Reuters corpus dataset includes a total of 8,213 file data, and each file data belongs to a category. In both data sets, stop words are removed, and each file is represented by a tfidf vector. The present invention sorts each word in the file according to their tfidf scores, and selects 1000 words with the largest score for each file as the feature of the file.
[0060] In order to objectively evaluate the performance of the algorithm of the present invention, the present invention uses the clustering method to evaluate in the selected test set, and uses Accuracy and NMI to evaluate the effect of the present invention, and Experimental solutions are carried out for the standards of dividing the files in the data set into 5, 7, a...
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 discloses a method for solving feature extraction tasks by using image regularization and data reconstruction. It mainly includes the following steps: 1) For a group of data points, construct its weight matrix and corresponding Laplacian matrix. 2) Randomly initialize the feature extraction matrix and reconstruction coefficient matrix, iteratively update the feature extraction matrix and reconstruction coefficient matrix, and obtain the final convergent feature extraction matrix as the basis for feature extraction. Compared with general item recommendation solutions, the present invention uses a method combining image regularization and data reconstruction, which can extract more effective data features. Compared with the traditional method, the effect obtained by the present invention in the problem of data feature extraction is better.
Description
technical field [0001] The invention relates to community question answering tasks, in particular to a method for solving feature extraction tasks by using image regularization and data reconstruction. Background technique [0002] In tasks such as data mining and information retrieval, the reduction of data dimensionality is a very important task. Reducing the dimensionality of data is of great significance for reducing the time and space consumption of processing data. overfitting phenomenon. The reduction of data dimension usually involves the problem of data feature extraction. For feature extraction, there are currently two methods: supervised feature extraction and unsupervised feature extraction. The supervised feature extraction method requires data to have label information, but currently The amount of data with label information is very small, so the present invention will adopt a non-supervised feature extraction method. [0003] The present invention will use t...
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): G06F16/2458
CPCG06F16/2465
Inventor 赵洲孟令涛高天祥何晓飞蔡登庄越挺
Owner ZHEJIANG UNIV