论文部分内容阅读
选用8个不同材料涂层压电石英晶体组成传感器阵列,所获频移数据用逐步判别(SDA)法解析,对小分子脂肪醇类与非醇类溶剂蒸气用二类判别法分析,找到由少于8个传感器组成阵列可达最佳判别能力,训练集回测正确率平均为84.8%,预测集识别正确率平均为87.3%。此外,该阵列用于识别酒、饮料,结果较佳。
Eight piezoelectric quartz crystal coated sensor arrays were selected. The frequency shift data obtained were analyzed by stepwise discrimination (SDA) method. Two kinds of discriminant methods were used to analyze the small molecule fatty alcohols and non-alcoholic solvent vapors. The array composed of less than 8 sensors achieves the best discrimination ability, the average accuracy of the training set back to test is 84.8%, and the recognition accuracy of the prediction set is 87.3% on average. In addition, the array is used to identify wine, drinks, the result is better.