应数学系郭玉坤副教授的邀请，长春理工大学理学院数学系尹伟石副教授将于近日来访公司并做报告，欢迎感兴趣的师生参加。

   题目：基于卷积神经网络和XGBOOST 的情感分析模型

   时间：2018年7月24日，上午10：00-11:00

   地点：格物楼503报告厅  

   摘要: Sentiment analysis is one of the important tasks in the text information mining of social networks. It has important theoretical and practical value in personalized recommendation, public opinion analysis and so on. However, sentiment analysis is not accurate when analyzing short texts and long texts. Combining the advantages of convolutional neural network model in feature extraction, a convolutional neural network hybrid model based on XGBoost's high-precision classification and fast computation is proposed. This model uses XGBoost to classify, which solves the problem of low accuracy caused by classification of softmax in convolutional neural networks (CNN). It has a good effect on the sentiment analysis of short texts and long texts. Finally, experiments were performed on data sets in multiple fields, and better results were obtained than the convolutional neural network model and the time recursive neural network.

报告人简介：尹伟石副教授，现为长春理工大学副教授。复旦大学数学学士，吉林大学计算数学硕士，吉林大学计算数学博士，主要从事科学计算、反问题、微分方程的数值解等方面的研究。他在声学逆散射计算，手性介质中声学散射，分数阶微分方程理论等领域做出了一批重要的工作。他合作主持国家自然科学基金1项，作为主要参与人参与国家自然科学基金2项，省部级项目多项，在《Advances in Mathematical Physics》、《Advances in Intelligent Systems Research》、《Advances in Computer Science Research》、《Advances in Modelling and Analysis》等杂志上发表高水平研究论文20余篇，出版学术专著2部。