The　application　of　traditional　collaborative　filtering　algorithm　on　large-scale　commercial　websites　is　very　mature.　However,　the　data　sparsity　and　extensibility　problems　that　occur　in　the　algorithm　affect　the　recommendation　accuracy　of　the　algorithm.　In　order　to　solve　this　problem,　a　SOM　clustering　collaborative　filtering　algorithm　based　on　singular　value　decomposition　is　proposed.　Firstly,　the　original　sparse　matrix　is　reduced　by　the　singular　value　decomposition,　and　the　items　are　evaluated　in　the　low-dimensional　space,　the　prediction　results　are　filled　in　the　original　matrix,　which　alleviates　the　problem　of　data　sparseness.　Then　use　SOM　to　cluster　the　users,　which　reduces　the　range　of　users　searching　for　neighbors　and　improves　the　scalability　of　the　algorithm.　The　experimental　results　on　MovieLens-100k　show　that　the　algorithm　can　effectively　improve　the　accuracy　of　the　recommendation.　©　2019　Association　for　Computing　Machinery.