ass日本风韵熟妇pics男人扒开女人屁屁桶到爽|扒开胸露出奶头亲吻视频|邻居少妇的诱惑|人人妻在线播放|日日摸夜夜摸狠狠摸婷婷|制服 丝袜 人妻|激情熟妇中文字幕|看黄色欧美特一级|日本av人妻系列|高潮对白av,丰满岳妇乱熟妇之荡,日本丰满熟妇乱又伦,日韩欧美一区二区三区在线

Crop disease diagnosis has accumulated a large number of electronic prescription data. It is an important practical problem that how to make secondary use of electronic prescription data to realize intelligent recommendation of crop disease prescription in the field of plant protection. A CDSSM-based crop disease prescription recommendation method was constructed to realize the diagnosis and prescription recommendation of multiple crop diseases. Based on the disease standard knowledge base, the crop disease prescription data were screened and expanded, and the standard prescription database was constructed combining with the domain knowledge. The CDSSM-based crop prescription recommendation model was constructed, semantic vector was generated according to text features, Cosine distances of semantic vectors were calculated, and prescription recommendation was completed with standard prescription database. The results were analyzed from four aspects of disease diagnosis, prescription recommendation, tomato disease prescription recommendation and influence of different inputs on prescription recommendation. The results were compared with models based on DSSM, DSSM-LSTM, Cosine, Jaccard and BM25. Combined with the actual application requirements, the mobile terminal oriented crop disease prescription recommendation application “Prescriptionist” was designed and constructed. The results showed that the accuracy of disease diagnosis of CDSSM was 71%, the accuracy of prescription recommendation was 82%, which were better than that of the other five crop disease prescription recommendation models. The recommendation accuracy of tomato disease prescription was higher. The CDSSM-based crop prescription recommendation model constructed can meet the practical application requirements, and also expand the disease types, which can be used as an efficient auxiliary tool for crop disease prescription recommendation.