基于随机森林算法的温度校准方法研究

Research on temperature calibration method based on random forest algorithm

基于热敏电阻NTC在生产过程、储存条件等客观因素的影响下会造成测量精度较低、准确性较低的情况,且传统校准方法受到热敏电阻非线性温度特性的影响,无法满足高精度和高准确性的校准需求,本文提出引进机器学习中的随机森林算法模型,在0~120℃的温度范围实验条件下,将测试数据分为训练集与测试集,有效验证了随机森林模型适用于NTC的温度校准,通过对模型的修改及优化参数,使温度校准精度达到-0.015~0.020℃。

Based on the fact that thermistor NTC can cause lower measurement precision and lower accuracy performance under the influence of objective factors such as production process and storage conditions,and that the traditional calibration method is affected by the nonlinear temperature characteristics of thermistor,which can not satisfy the calibration results of high precision and accuracy,this paper proposes to introduce the random forest algorithm model in machine learning,and in the experimental conditions of the temperature range of 0-120℃,the test data are divided into training set and test set,and the random forest model is effectively verified to be suitable for the temperature calibration of NTC,and the temperature calibration accuracy reaches-0.015-0.02℃by modifying the model and optimizing the parameters.