【摘要】 【目的/意义】微博舆情对社会各领域的影响与日俱增,但由于其影响因素众多,呈现出非线性且复杂的变化。因此,如何快速、准确地预测其发展趋势是一个很有价值的研究课题。【方法/过程】以微博话题的博文总数作为微博话题发展趋势的量化指标,考虑话题发展的复杂性和非线性的特点,采用模糊神经网络来预测微博话题的发展趋势。并通过改进的粒子群优化算法对模糊神经网络的参数进行优化以更好的发挥模糊神经网络在处理非线性、模糊性等复杂问题上的优越性。【结果/结论】通过对新浪微博数据集的对比实验,验证了本文所提方法的有效性和准确性。本文方法有效解决了微博舆情趋势预测中遇到的模型参数复杂、易陷入局部最优的问题,提高了微博舆情发展趋势预测的准确性。更多还原

【Abstract】 【Purpose/significance】Microblog public opinion is playing an significant role in many society fields in recentyears. But due to so many affect factors,it presents a nonlinear and complex changes. How to fleetly and accurately predictits development trend has become a meaningful research.【Method/process】Taking all of the microblog for certain topics asa quantitative indicator of the development trend of microblog, considering the complexity and nonlinearity of topic change,the fuzzy neural network is used to predict the trend of microblog topic. And improved Particle Swarm Optimization Algo-rithm is used to optimize the parameters of fuzzy neural network to better play the advantages of fuzzy neural network indealing with complex problems such as nonlinear, fuzzy and so on.【Result/conclusion】The result of simulated experimen-tal analysis on sina Wei Bo data shows that it is efficient and accurate for the method suggested in this paper. This methodcan solve problems of complex model parameters and easily to run into partial optimization in microblog public opiniontrend prediction, and improve the accuracy of trend predicting of microblog public opinion.更多还原