Improvement of Chromatographic Peaks Qualitative Analysis for Power Transformer Base on Decision Tree

The decision tree applied to the chromatographic peak characterization of power transformers in this task. The follows parameters taken as the characteristic attributes of chromatographic peak identification that include peak height, peak width, peak area and peak position. Dichotomy adopted to discretize the continuous attributes to select the nodes by decision tree. From the algorithm, not only ones obtain the adaptive threshold of the characteristic attributes and achieve the correct classification of single effective peaks, but also avoid the errors caused by the simultaneous decision of seven component peaks. The test results shown that the improved algorithm has the advantages of simple principle, good peak drift resistance and false peaks eliminated.