Intelligent recognition method of toxic gas in electronic nasal chamber based on semi-supervised learning
A semi-supervised learning and intelligent recognition technology, applied in the field of classification and recognition, can solve the problems of inability to meet application requirements, poor learning effect of test samples, and low classification accuracy, achieve strong ability to learn smell patterns, and improve the scale of basic classifiers Effect
Inactive Publication Date: 2018-01-09
SOUTHWEST UNIV
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] However, there are deficiencies in the proposed semi-supervised learning techniques: first, a large part of the semi-supervised learning techniques are aimed at binary classification problems, and the types of indoor poisonous gases are far more than two, so they cannot meet the application requirements; In the semi-supervised learning algorithm of classification, the size of the classifier is limited, resulting in poor learning effect on test samples and low classification accuracy
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
[0022] The specific implementation manner and working principle of the present invention will be further described in detail below.
[0023] A semi-supervised learning-based intelligent poison gas identification method in an electronic nasal chamber, which is carried out according to the following steps:
[0024] Step 1: Obtain the poisonous gas sample data set L with known labels and the poisonous gas sample data set U with unknown labels. The number of preset basic classifiers is M=3, and the current number of training times is t;
[0025] Step 2: Randomly generate M subsets L of equal size from the toxic gas sample data set L with known labels i to train each base classifier c i , i=1~M;
[0026] Step 3: Use each basic classifier trained in step 2 to classify and identify the poisonous gas sample data set L with known labels, obtain the initial recognition rate of each classifier, and use the simple voting method to evaluate the discrimination results of all classifiers ...
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 present invention discloses an intelligent identification method of electronic nasal poison gas based on semi -supervised learning. The sample training of the known labeling gas sample data set data is used to train each basic classifier.As the main classifier, the unknown label sample data set U is classified through the main classifier, and the other basic classifiers predict the label of the data in the sample data set U. In the voting resultsWhen the number of votes of a data label exceeds the pre -set threshold, the sample data will be used to re -train the classifier with its original data set L, and finally use the number of classifiers to determine the system's system.Whether the recognition rate has reached the optimal, so that the trained classifier not only has more basic classifiers, but also has the ability to learn odor mode in the unknown sample.
Description
technical field [0001] The invention relates to a classification and recognition technology in electronic nose signal processing, in particular to an intelligent recognition method of poisonous gas in an electronic nasal chamber based on semi-supervised learning. Background technique [0002] At present, for indoor toxic gas detection, in order to ensure the correctness of the detection results, the electronic nose system used needs to use a large number of learning samples in the training stage. It is higher than the electronic nose trained on unlabeled data, but unlabeled data is easier to obtain than labeled data. [0003] Therefore, someone proposed a semi-supervised learning technology, which can help the electronic nose not only learn related patterns from training samples, but also learn related knowledge from samples with unknown labels, so as to achieve continuous learning of a certain odor pattern. Until the recognition rate no longer occurs any improvement. [0...
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): G06K9/62
CPCG06F18/2411G06F18/214
Inventor 贾鹏飞段书凯王丽丹葛灵普黄泰来朱赛克高锦程陈祥宇闫嘉
Owner SOUTHWEST UNIV