With　the　development　of　the　Internet,　the　information　in　the　network　is　growing　explosively,　and　the　information　surplus　brings　a　great　burden　to　people　in　screening　and　selecting　information.　Collaborative　filtering　recommendation　can　effectively　realize　information　filtering.　There　are　many　applications　in　various　scenarios.　Currently,　collaborative　filtering　recommendation　methods　can　be　roughly　divided　into　two　types:　explicit　feedback　based　recommendation　and　implicit　feedback　based　recommendation.　In　recent　years,　algorithms　for　implicit　feedback　recommendation　have　developed　rapidly,　and　many　excellent　algorithms　have　emerged,　the　performance　of　recommendation　algorithm　based　on　explicit　feedback　still　needs　to　be　improved.　This　paper　analyzes　and　models　human　cognitive　behavior,　and　proposes　a　subjective　and　objective　evaluation　transformation　model.　Based　on　this　model,　a　recommendation　algorithm　based　on　explicit　feedback　is　proposed.　Simulation　results　show　that　the　algorithm　has　good　performance　in　recommendation　accuracy.　©　2020　IEEE.