The　competitive　landscape　of　multiple　e-commerce　platforms　and　the　vast　amount　of　product　reviews　associated　with　these　platforms　have　supported　both　consumers'　online　shopping　decision　making　and　also　served　as　a　reference　for　product　attribute　performance　improvement.　This　paper　proposes　a　sentiment-driven　fuzzy　cloud　multi-criteria　model　for　online　product　ranking　and　performance　to　provide　purchase　recommendations.　In　this　novel　model,　Bi-directional　Long　Short-Term　Memory　Network-Conditional　Random　Fields　(BiLSTM-CRF),　sentiment　analysis,　and　K-means　clustering　are　first　integrated　to　mine　product　attributes　and　compute　sentiment　values　based　on　reviews　from　various　platforms.　Next,　considering　the　confidence　of　the　sentiment　value,　the　cloud　model　is　combined　with　q-rung　orthopair　fuzzy　sets　to　define　the　new　concept　of　the　q-rung　orthopair　fuzzy　cloud　(q-ROFC)　and　the　interaction　operational　laws　between　q-ROFCs　are　given.　The　sentiment　values　of　each　product　attribute　from　different　platforms　are　cross-combined　and　transformed　into　a　type　of　q-ROFC,　while　multiple　interactive　information　matrices　are　established.　To　investigate　the　correlation　among　homogeneous　attributes,　the　q-ROFC　interaction　weighted　partitioned　Maclaurin　Symmetric　mean　operator　is　proposed.　Finally,　we　provide　real-world　examples　of　online　mobile　phone　ranking　and　attribute　performance　evaluation.　The　results　show　that　our　proposed　method　offers　significant　advantages　in　dealing　with　customer　purchase　decisions　for　online　products　and　problems　with　performance　direction　identification.　Managerial　implications　are　discussed.　IEEE