Factor analysis and SOM network-based project schedule risk prediction method

The invention discloses a factor analysis and SOM network-based project schedule risk prediction method. According to the method, factor analysis and clustering analysis are combined, a result is output to SOM, while observation data is not directly trained in an artificial neural network, so that the load of the neural network is reduced greatly by a statistic tool, the operating efficiency of SOM is improved, and the neural network avoids non-converging due to massive operating or does not get in a trouble of local minimum; the SOM network with high mode identification power predicts the sample, trains fewer times, is clear in classification, and is highly adaptable to human factors in actual production. Particularly when an array of projects comes, data volume is large, and the data dimension is high, the prediction system fully utilizes the big data statistics analysis advantage of MINITAB and the intelligent calculation advantage of MATLAB. 本发明公开了种基于因子分析和SOM网络的项目进度风险预测方法，用因子分析结合聚类分析，将其结果输给SOM，而非直接将观测数据赋给