基于机器学习的科研期刊分类方法
A Machine Learning Based Approach to Categorize Research Journals
摘要：
在这个现代技术时代，研究期刊的分类和排名越来越受到研究人员和科学家的欢迎。它对他们的研究成果在高质量期刊上发表起着重要作用。虽然目前学术界对期刊分类和排序的研究很多，但是利用合适的机器学习技术对期刊进行分类和预测的研究还很缺乏。这项工作旨在对各种学术研究期刊进行分类和预测。这项工作提出了一个混合预测模型，包括五个步骤。第一步是准备具有20个特征的数据集。第二步是对数据集进行预处理。第三步是应用适当的聚类算法进行分类。第四步是应用适当的特征选择技术得到有效的特征子集。第五步，采用集成与非集成相结合的方法对模型进行训练。该模型在一个完整的特征集上进行训练，并通过应用三种特征选择技术来获得所选的特征子集。在模型训练之后，预测结果在精确度、召回率和准确度方面进行评估。研究结果可供研究者与实务工作者预测期刊类别之参考。
英文摘要：
In this modern technological era, categorization and ranking of research journals is gaining popularity among researchers and scientists. It plays a significant role for publication of their research findings in a quality journal. Although, many research works exist on journal categorization and ranking, however, there is a lack of research works to categorize and predict the journals using suitable machine learning techniques. This work aims to categorize and predict various academic research journals. This work suggests a hybrid predictive model comprising of five steps. The first step is to prepare the dataset with twenty features. The second step is to pre-process the dataset. The third step is to apply an appropriate clustering algorithm for categorization. The fourth step is to apply appropriate feature selection techniques to get an effective subset of features. The fifth step involves some ensemble plus non ensemble methods to train the model. The model is trained on a full set of features, and a selected subset of features is obtained by applying three feature selection techniques. After model training, the prediction results are evaluated in terms of precision, recall, and accuracy. The results can help the researchers and the practitioners in predicting the journal category.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
