Abstract: This paper constructs a transaction bipartite network between patent technology and subject. Considering the influence of technology preferences, degree of common neighbors and transaction frequency on transaction opportunities, this paper revises the pattern weight covered by links and proposes the technology transaction opportunities prediction model based on multidimensional projection of bipartite network. Through the empirical study of patent technology transaction data in the field of new energy and the comparison with the traditional bipartite network link prediction method, this paper concludes that the hit rate of the method proposed in this paper is significantly higher than that of the traditional method, and the accuracy and robustness of this method are verified by randomly selected 10 groups of data sets for comparison. In addition, the empirical study shows that the technology transaction within large enterprise groups is active but the external radiation is weak. Universities and scientific research institutes are less involved in transaction. The historical transaction can enhance the trust relationship and cooperation experience among the subjects, which has an important impact on the transaction prediction, and the technology and region preferences of subjects transaction also contribute to the improvement of prediction accuracy. The ideas and method proposed in this paper can not only realize the recommendation of patent technology transaction, but also carry out trading partner recommendation based on technology recommendation, and predict the hot areas of future transaction, which can provide decision support for promoting the development of technology market.

Key words: bipartite network, multidimensional projection, technology transaction opportunity, link prediction