The invention relates to an independent
station commodity query method and device, equipment and a medium, and the method comprises the steps: obtaining a commodity query word expressed in a non-central language, determining whether the commodity query word is a
polysemy word or not according to a
polysemy word corpus, and calling a central
language translation model to translate a plurality of translation words expressed in a central language when the commodity query word is the
polysemy word; inputting a commodity query word and a plurality of translation words thereof into a multi-category prediction model, predicting the confidence coefficient of each category mapped to the independent
station commodity category
system, selecting a plurality of categories with high confidence coefficients as candidate categories, and converting the confidence coefficients into weights; corresponding to candidate commodities in a commodity candidate
list recalled according to the plurality of translation words, weighting predicted scores of target categories to which the candidate commodities belong in a commodity category
system by using weights to determine sorting scores; and selecting target commodities from the candidate commodities according to the sorting
score to form a commodity recommendation
list. Therefore, the commodity better matched with the intention of the user can be obtained by compatibility with
ambiguity during query word translation.