New Word Discovery Algorithm for Power Marketing Audit Combining Adjacency Entropy and Random Forest

Due to the large amount of power marketing audit data, the accuracy of new word recognition is not high. To this end, a new word discovery algorithm for power marketing audits combining adjacency entropy and random forest is designed. The power marketing audit words are preprocessed through the steps of removing format tags, word division, stemming, and removing stop words. The power marketing audit words are structured through feature vectors, and the random forest algorithm is used to classify power marketing audit words. The neighborhood entropy is used to filter the candidate words to get new candidate words for power marketing audit. Through comparative experiments, it is compared with two traditional algorithms. The experimental results show that the proposed new word discovery algorithm for power marketing audits combining adjacency entropy and random forest has better word segmentation effect, which is 10.603% higher than traditional algorithm 1 and 10.578% higher than traditional new word discovery algorithm 2.