基于社会媒体的产品销量预测技术

项目名称： 基于社会媒体的产品销量预测技术

项目编号： No.61472107

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 计算机科学学科

项目作者： 刘挺

作者单位： 哈尔滨工业大学

项目金额： 80万元

中文摘要： 社会媒体已经迅速发展成为具有重大影响力的新媒体，并在很多领域表现出了惊人的预测能力。基于社会媒体的产品销量预测，通过对社会媒体数据的挖掘与分析，聚集大众的群体智慧，运用科学的知识、方法和手段，对产品销量未来发展趋势和状态做出科学的估计和评价。本项目旨在构建基于社会媒体的统一通用的产品销量预测模型，其主要特点和创新点体现在如下三方面：1）本项目面向产品销量预测问题提出消费意图挖掘任务，相对于广泛应用于产品销量预测中的用户口碑分析，消费意图能够更加直接地反映出产品销量的未来走势；2）本项目基于深度神经网络建模学习现实世界中的事件与产品销量之间隐含关系，将当前发生的事件对产品销量可能产生的影响直观地呈现出来；3）本项目提出基于社会媒体的预测模型将基于消费意图与基于事件的产品销量预测相结合，进而将相关关系与因果关系、线性模型与非线性模型整合到统一的预测模型当中，为产品销量预测提供一体化的解决思路。

中文关键词： 社会计算；自然语言处理；文本挖掘；信息抽取；人工智能

英文摘要： The past few years have brought huge growth in social media, and it has become a new media that rivals the traditional media. In nowadays, social media has shown its power on prediction in various domains which motivates us to conduct product sales prediction by exploiting social media content. Social media based prediction refers to people use scientific knowledge, methods and means to make a scientific estimation and evaluation for the future development trends and statements based on data mining and analysis on social media. This project intends to build a unified social media based product sales prediction model. The key characteristics and contributions are as follows. (1) In this project, we propose the task of consumption intention mining for solving the problem of product sales prediction. Consumption intention mining can more directly reflect the trends of product sales compared to the sentiment analysis which is widely used in tradional prediction model; (2) Based on the deep neural network, this project models the hidden and complex relationship between events in the real world and the product sales. The model may directly show the impact of events to product sales; (3) In this project, we focus on building social media based prediction model which integrates the consumption intention and events to predict product sales, and further combines correlation and causality, linear model and nonlinear model in our prediction model, which can provide a unified research solution for the problem of product sales prediction.

英文关键词： Social Computing;Nature Language Processing;Text Mining;Information Extraction;Artificial Intelligence