谣言判别器的设计与实现

Design and implementation of rumor discriminator

在互联网时代,信息传播的速度和范围都得到了极大的提升,使得各种信息能够在极短的时间内迅速扩散至广泛的受众,然而这也带来了一些挑战,谣言的传播是其中之一,自动谣言判别可以大大降低谣言传播率。本文构建了中文谣言数据集,设计实现了谣言自动检测的小程序:用户输入一段言论,小程序的前端页面将数据传送至后端,基于卷积神经网络的模型进行自然语言处理,模型采用深度学习算法对该言论进行语义建模和分类,最终得出这段话是谣言的概率并返回给用户。

In the Internet era,the speed and scope of information dissemination have been greatly improved,which enables all kinds of information to rapidly spread to a wide range of audiences in a very short period of time,however,this also brings some challenges,the spread of rumors is one of them,and automatic rumor discernment can greatly reduce the rate of rumor dissemination.In this paper,we constructed a Chinese rumor dataset,and designed and implemented an applet for automatic rumor detection:the user inputs a piece of speech,the front-end page of the applet transmits the data to the back-end,the model based on convolutional neural network carries out the natural language processing,and the model adopts a deep learning algorithm to semantically model and classify the speech,and ultimately arrives at the probability of the piece of speech being a rumor and returns it to the user.