Market transaction data has the characteristics of strong professionalism,high semantic complexity,and mixed subjective and objective information,which makes the intelligence elements for the technology market complex.It is necessary to build a deep and fine-grained intelligence scanning system to support scientific and technological innovation and industrial development research.By this,the research uses a comprehensive evaluation system,deep learning model and management principles to first propose a hot spot scanning model for technology market transactions.Firstly,the entropy weight method is combined with complex and multi-dimensional transaction data to construct a comprehensive evaluation model of transaction heat;Secondly,the BiLSTM+LAC joint entity recognition model is constructed to automatically identify transaction objects from numerous contextual and interdependent text data;Finally,the transaction hotspots scanning is carried out according to the Pareto Principle.Through the above model,the transaction hotspots in the medical device field of the technology market are scanned,and the results are consistent with the data survey.Therefore,the scanning model proposed in this paper can reasonably identify the hot spots of technology market transactions,which is of great significance to the technology market construction and industrial layout research.