基于极速神经网络的协作过滤方法研究

Efficient collaborative filter using extreme learning machine

协作过滤是一种有效的个性化推荐技术,针对该技术随着用户和资源的增多,数据的高维稀疏特性严重导致推荐质量的下降和计算速度减慢的问题,研究并实现了一种基于极速神经网络的协作过滤方法。采用主成分分析解决数据高维稀疏性问题,采用极速神经网络技术解决计算速度慢的问题。实验结果表明,该方法具有良好的泛化性能和学习速度,能很好的满足个性化资源推荐的需求。

Collaborative filtering as an effective technique for personalized recommendation has been widely applied in many fields,but with the increase of users and resources,high-dimensionality and scarcity has seriously degrade the recommendation quality and slow down the calculation speed.According to this problem,collaborative filtering based on extreme learning machine has been proposed and implemented.Principal component analysis is used to solve high-dimensional sparse problem,while extreme learning machine is applied to solve the problem of slow speed.Experimental results have shown that the proposed method has good generalization performance and learning speed,and it can meet the demand for personalized recommendation well.