To　effectively　address　challenges　that　stem　from　e-commerce,　it　is　crucial　to　harness　diverse　review　data　from　e-commerce　platforms.　These　data　support　consumers　in　making　informed　purchase　decisions　and　aid　manufacturers　in　optimizing　product　attributes.　Incorporating　sentiment　data　from　heterogeneous　reviews　across　different　time　periods　into　a　decision-making　framework　is　a　pivotal　consideration　in　purchase　decisions　and　product　design.　The　goal　of　the　study　is　to　establish　an　online　product　decision　support　method　grounded　in　consumer　irrational　behavior　and　segmented　reviews　over　time.　It　aims　to　offer　users　reliable　and　consistent　outcomes　when　making　personalized　purchase　decisions.　The　probabilistic　linguistic　term　set　is　employed　to　represent　consumer　sentiments　with　varying　degrees　of　granularity　across　different　time　periods.　Subsequently,　stochastic　sampling　is　utilized　to　simulate　the　decision-making　process　of　individual　consumers.　Regret　theory　is　then　applied　to　analyze　consumers＇　irrational　psychological　behavior.　Building　upon　heterogeneous　data　gathered　from　e-commerce　platforms,　including　review　ratings,　likes,　and　follow-up　reviews,　a　multiperiod　group　decision　approach　based　on　maximum　similarity　and　review　helpfulness　is　proposed.　This　decision-making　method　is　advanced　through　a　decomposition-aggregation　process,　safeguarding　against　information　distortion　and　ensuring　result　reliability.　This　method　provides　consumers　with　product　selection　solutions　across　the　temporal　dimension　and　serves　as　a　theoretical　compass　for　manufacturers　and　sellers　seeking　product　enhancement　and　sales　optimization.　IEEE