Development and Verification of Multi-class Data in Artificial Intelligence Model Group Based on Cloud Computing

Artificial intelligence (AI) has been widely studied. Due to the increasingly wide application of multi-class data, with the characteristics of rapid arrival, rapid change, huge quantity, and potential infinity, the original classification algorithm for the static database cannot adapt to the reality of multi-class data. In order to achieve reliable and accurate data dependency analysis by loop-level parallelism detection in large programs, this paper firstly designs and implements a parallel Naive Bayes (NB) multiclass data analysis algorithm. Then, we adjust the number and weight of classifiers as needed to maintain high classification accuracy and reduce the cost. Meanwhile, this paper establishes an optimal complexity model by realizing the self-organizing control of the data mining process. The results show that the parallel NB classification algorithm can bring excellent performance improvement and good scalability and can meet individuals’ performance requirements for massive data classification to a certain extent.