US 12,235,003 B2
Method and system for predicting response and behavior on chats
Jay Ketankumar Trivedi, Gujarat (IN); Shashi Singh, Uttarakhand (IN); Pankaj Tanwar, Himachal Pradesh (IN); Sonal Jain, Gujarat (IN); and Rohit Kumar, Bihar (IN)
Assigned to Samsung Electronics Co., Ltd., Suwon-si (KR)
Filed by SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR)
Filed on Jul. 28, 2022, as Appl. No. 17/875,567.
Application 17/875,567 is a continuation of application No. PCT/KR2022/010121, filed on Jul. 12, 2022.
Claims priority of application No. 202111042752 (IN), filed on Sep. 21, 2021.
Prior Publication US 2023/0107375 A1, Apr. 6, 2023
Int. Cl. G06F 15/16 (2006.01); F24F 11/46 (2018.01); F24F 11/56 (2018.01); F24F 11/80 (2018.01); H04L 51/02 (2022.01); H04L 51/216 (2022.01); F24F 110/10 (2018.01)
CPC F24F 11/46 (2018.01) [F24F 11/56 (2018.01); F24F 11/80 (2018.01); H04L 51/02 (2013.01); H04L 51/216 (2022.05); F24F 2110/10 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A method for predicting a response of a user in a multi-user environment, comprising:
displaying, on a display screen, a user interface (UI) for inputting a message corresponding to one of at least one received message or a group of received messages from one or more contacts in one of a single window environment or a multi-window environment;
generating, by a first processing engine, a list of correlated candidate messages associated with an input message based on one of the at least one received message or the group of received messages;
predicting, by a second processing engine, a set of correlated candidate messages relevant to the input message from the list of correlated candidate messages;
determining, by the second processing engine, a plurality of weights for each of the correlated candidate messages in the generated list of correlated candidate messages based on subject of the input message, at least one of query satisfaction parameters or behavioral parameters, and at least one of user's potential to be a candidate for reply or a plurality of time attributes, and a message correlation weight for each of the correlated candidate messages in the generated list of correlated candidate messages by combining the determined plurality of weights for the respective correlated candidate message;
generating, by a third processing engine, a priority list of candidate messages including the predicted set of correlated candidate messages based on the message correlation weights of the correlated candidate messages; and
displaying, on the display screen, at least one of an appropriate candidate message or a list of the appropriate candidate messages based on a priority associated with the appropriate candidate messages in the generated priority list of candidate messages,
wherein the plurality of weights includes subject weights, at least one of query satisfaction weights or behavioral weights, and at least one of candidate weights or time-relevant weights corresponding to the correlated candidate messages.