基于ID3决策树改进算法的稿刊推荐研究

Research on Manuscript Publication Recommendation Based on UsingImproved ID3 Decision Tree Algorithm

在科研工作者的科研活动中,发表论文是其中非常重要的部分。论文承载着科研工作者的研究成果,只有发表在出版物上,才能得到世人的认可。现有稿刊推荐研究的推荐结果少,而且推荐结果的准确度不高,投稿人很难从推荐结果中发现适合投稿的期刊。鉴于此,提出利用改进的ID3决策树算法对期刊与稿件相关度进行分类建模,利用得到的分类规则为投稿人推荐合适期刊的方法。实验结果表明,基于ID3决策树改进算法的稿刊推荐方法推荐准确率较高。

In the research activities of scientific research workers, to publish papers is one of the most important parts. The papers carry the research results of scientific research workers. Nowadays, the results of the research manuscript-publication recommendation are less and the accuracy is low. It is difficult for the contributors to find the journals suitable for submission from the recommendation results. Aiming at this phenomenon, the ID3 decision tree algorithm is proposed to classify the relevance of journals and manuscripts, and use the obtained classification rules to recommend the related journals to manuscripts. Firstly, the shortcomings of the traditional ID3 decision tree algorithm are summarized, and then the process of improving ID3 decision tree algorithm by multi-value logic is expounded. Finally, the classification rule extracted by the generated decision tree model is used to the manuscript-publication recommendation. The experimental results show that the proposed method based on the improved algorithm of ID3 decision tree is not only recommend tbe related journals, but also has high accuracy.