This　research　utilizes　data　mining　and　text　analysis　methods　to　extract　investor　attention,　emotion,　and　opinion　convergence　from　stock　forums.　A　Restricted　Reverse　MIDAS　model　is　constructed　to　explore　the　dynamic　effects　of　real　-time　public　opinion　on　intraday　high-frequency　stock　index　and　trading　volume.　The　empirical　results　demonstrate　a　positive　impact　of　forum　public　opinion　on　high-frequency　stock　index　and　trading　volume,　with　a　time　attenuation　effect　observed.　Specifically,　midday　public　opinion　significantly　influences　the　afternoon　highfrequency　stock　index,　exhibiting　an　initial　increase　and　subsequent　decrease　in　investor　sentiment,　while　investor　attention　and　opinion　convergence　exhibit　a　decreasing　trend.　Nighttime　attention　and　sentiment　primarily　affect　the　next　morning＇s　high-frequency　stock　index.　Moreover,　the　real　-time　emotional　impact　on　the　stock　market　varies　across　different　environments.