Construction of user Portrait System Based on information warehouse under big data

With the continuous development of Internet technology, the scale of data volume shows an explosive growth trend.In this case, the traditional user portrait system will have problems such as insufficient data storage and slow calculation. Based on this, this paper designs a user portrait technology based on information warehouse under big data. Through the data acquisition module, the distributed information collection is carried out to complete data acquisition, data cleaning and zoning. Through the data warehouse hierarchical operation, the data is accumulated and summarized, and complex tasks are resolved into multiple simple tasks. It is convenient to locate problems, improve the reusability of calculation results, and decouple original data and statistical data. Through user portrait management platform module, the labels are classified, customized. For the labels of statistical classes are distributed and aggregated through spark. For the labels of mining classes, K-means algorithm is used to build a multi-level user portrait model. The storage of portrait data and the screening of multi label combination are realized through the user clustering module. The experimental results show that the system is suitable for the construction of user portrait in the case of large amount of data. The advantages of the system are lower delay and stable accuracy.