基于机器学习的军事装备知识分类方法

Knowledge Classification Method for Military Equipment Based on Machine Learning

基于开源获取的军事百科知识提取知识中关键特征并赋予特征权重,分别以词频-逆文档频率(TF-IDF)法和词向量(Word2Vec)作为文本表征手段,采用K最近邻(KNN)、支持向量机(SVM)、神经网络及其他机器学习算法开展军事装备知识分类研究。提出了装备知识大类(装备、地点和部队等)、装备目录层级小类2级分类模式,取得了较好的分类结果;比较了各算法的优劣,有助于形成更高效、准确的军事装备知识模型,可支撑军事装备知识图谱的构建和应用。

Based on the military encyclopedia knowledge acquired from open source,the key features of knowledge are extracted and the feature weights are given.Taken the term frequency-inverse document frequency(TF-IDF)algorithm and Word2 Vec as the text representation means,using K-nearest neighbor(KNN),support vector machine(SVM),neural network and other machine learning algorithms,the knowledge classification research of military equipment is carried out.The two-level classification mode including the first type of the equipment knowledge and the second type of the equipment catalog level are presented,and good classification results are achieved.The advantages and disadvantages of each algorithm are compared.It’s helpful to form a more efficient and more accurate military equipment knowledge model,and it can support the construction and application of the military equipment knowledge graph.