In today’s rapid development of information technology and big data technology, consumer behavior is undergoing a profound transformation. This study focuses on the decision-making stage of consumer journey, selects indicators based on webpage click stream data, improves the K-means algorithm, and realizes the identification of consumer journey nodes using the binary K-means algorithm. Based on the review recommendation scenario, from the perspective of consumer decision-making journey, we introduce the “attention-attitude-understanding-purchase intention” stage-based decision-making model, apply it to the model design of deep learning, and combine the attention mechanism and co-attention mechanism to propose a product recommendation method based on online reviews. The results show that consumers in clusters 1-4 are in the consumer journey nodes of attention, understanding, attitude, and purchase intention, respectively. The product recommendation model exhibits better recommendation accuracy and time efficiency, with accuracy improved by 18.72%~67.12% and time reduced by 8.39%~62.03% over the comparison method. This paper realizes the innovation of deep learning method with the support of consumer behavior theory, and improves the methodological technical support for accurate online marketing strategy.
1970-2025 CP (Manitoba, Canada) unless otherwise stated.