Electronic medical record characteristic selection method based on intra-class distribution and inter-class distribution of words

The invention relates to an electronic medical record characteristic selection method based on the intra-class distribution and the inter-class distribution of words. The method comprises the following steps: firstly, constructing a disease and drug dictionary, secondly, summarizing the three characteristics of the symptom word distribution in Chinese medical records, preprocessing the collected patient medical records, and storing the word segmentation results in a database. According to the invention, the representative degree of the words for different diseases is calculated according to the distribution condition of the key symptom words in the diseases and the distribution condition among the diseases, and moreover, when the distribution condition among the diseases is considered, andthe influence of correlation between diseases on characteristic selection is analyzed, and the characteristics of the diseases with high relevance are ensured to be similar to the greatest extent when the characteristics are