The　sentiment　of　the　stock　market　can　reflect　the　behavior　of　investors　to　a　certain　extent　and　influence　their　investment　decisions.　As　a　kind　of　unstructured　data,　market　news　can　reflect　and　guide　the　overall　environmental　sentiment　of　the　market.　Together　with　stock　prices,　it　can　become　a　crucial　market　reference　data,　which　can　effectively　help　investors＇　investment　decisions.　This　paper　proposes　a　vectoring　method　that　can　accurately　and　quickly　establish　a　multiple-bit　emotional　feature　for　massive　news　data.　It　uses　a　support　Victor　Machine　(SVM)　model　to　predict　the　impact　of　financial　news　on　the　stock　market　and　uses　bootstrap　to　alleviate　the　problem　of　over-fitting.　The　results　of　the　experiments　conducted　on　the　Shanghai　and　Shenzhen　stock　indexes　indicate　that　the　proposed　method　can　improve　the　prediction　accuracy　by　about　8%　compared　with　the　traditional　model,　and　obtained　an　excess　return　of　6.52%　in　the　three-month　backtest,　which　proves　its　effectiveness.　©　2021,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Switzerland　AG.