Sensor array optimization method for Carya cathayensis freshness detection

Abstract

The invention discloses a sensor array optimization method for Carya cathayensis freshness detection. The method includes: firstly selecting gas sensitive sensors according to a GC-MS analysis result of Carya cathayensis volatile gas; then selecting sensors to compose an initial array according to the response signals of sensor arrays in the volatile gas; extracting and retaining an average differential value, a stable value and an area value to compose an initial characteristic matrix; then taking characteristic values as the object, conducting mean value analysis on the characteristic values of Carya cathayensis with different freshness; then conducting variation coefficient analysis on the characteristic values; the taking the square Euclidean distance as metric, conducting clustering analysis on the characteristic values, and carrying out correlation analysis on various characteristic values; and finally conducting variance inflation factor analysis on a characteristic matrix to obtain a final optimized characteristic matrix and sensor array. The method provided by the invention performs optimization according to non-searching feature selection strategy, to optimize, picks out characteristic values with more effective information as far as possible, and is conducive to improving the accuracy of pattern recognition.

Description

technical field [0001] The invention relates to the field of food detection, in particular to a sensor array optimization method for detection of pecan freshness. Background technique [0002] Pecan is popular among consumers for its rich nutrition and unique taste. However, its internal oil is easily oxidized and rancid due to environmental factors such as moisture, light, and oxygen, thereby affecting its taste. Therefore, testing the freshness of pecans has certain practical significance. [0003] The smell of hickory is the volatile component emitted by its inner kernel and shell, and the change of external smell reflects the change of its internal freshness to a certain extent. The existing pecan freshness detection methods mainly include sensory evaluation, microbial detection, physical and chemical index determination, etc. However, the above-mentioned methods have the disadvantages of subjectivity, high detection cost and time-consuming, so that they cannot be well ...

AI Technical Summary

Problems solved by technology

Although the above examples have achieved good discrimination and prediction results in practical applications, since different eigenvalues ​​of different sensors may contain valid information in the process of pattern recognition, and their contribution rates to pattern recognition are not the same, Therefore, the sensor array obtained in the above example is often a better optimized array, and there is still room for optimization.

Images