Characteristic spectrum area selection method for near infrared spectrum

Abstract

The invention provides a characteristic spectrum area selection method for a near infrared spectrum. The characteristic spectrum area selection method comprises the following steps of: applying a Monte Carlo probability selection combined ant colony optimization algorithm to a characteristic spectrum area selection problem of the near infrared spectrum; setting a dynamic section range and initializing algorithm parameters to obtain each spectrum section of an object to be taken as an equivalent searching point; establishing a partial least square analyzing model by taking the quality or characteristics of the object to be detected as a standard reference; predicating a root-mean-square error by the model to repeatedly carry out weighting calculation to update pheromone vectors according to the predicated root-mean-square error; carrying out iterative computation and searching to obtain the optimal characteristic spectrum area of the near infrared spectrum; and carrying out multiple circulating calculation and automatically judging to obtain the optimal characteristic spectrum area of the near infrared spectrum. The characteristic spectrum area selection method disclosed by the invention combines the wholeness of Monte Carlo probability selection and ant colony optimization algorithm positive feedback so as to effectively avoid the disadvantages of experience selection in a modeling process and data redundancy of all selections, rapidly obtain a global optimum characteristic spectrum area, and improve the precision and the stability of modeling.

Description

technical field [0001] The invention relates to the technical field of near-infrared spectrum analysis, in particular to a method for selecting a near-infrared spectrum characteristic spectrum area based on a Monte Carlo-ant colony optimization algorithm. Background technique [0002] With the development of near-infrared spectroscopy and stoichiometric methods, near-infrared spectroscopy has been applied in various fields of national economic development. However, since the data collected by the instrument contains other irrelevant information and noise, such as electrical noise, sample background, etc., in addition to the sample's own information, it is difficult to completely eliminate these information in the preprocessing; secondly, the information of samples in some areas is very weak , lack of correlation with the composition or properties of the sample. If all these data are involved in modeling, not only will the calculation amount be large, the model will be compl...

AI Technical Summary

Problems solved by technology

However, since the data collected by the instrument contains other irrelevant information and noise, such as electrical noise, sample background, etc., in addition to the sample's own information, it is difficult to completely eliminate these information in the preprocessing; secondly, the information of samples in some areas is very weak , lack of correlation with the composition or properties of the sample

If all these data are involved in modeling, not only the amount of calculation is large, the model is complex, but the accuracy is also affected

Images