Mobile searches utilizing a query-goal-mission structure

A client device for improving mobile searches, the client device comprising:

• a processor;
• a display;
• and
• a memory storing instructions that, when executed on the processor, provide a system comprising:
• a mobile search component configured to:
• identify a feature, generated by a user within a time period on at least one client device, based upon an identification of the user;
• evaluate the feature to identify a goal associated with the feature;
• identify a mission associated with the goal;
• identify a second goal associated with the mission, the second goal having goal similarity within a similarity range of the goal;
• identify a query associated with the second goal, the query having a weighted common goal similarity within a goal similarity range of the feature;
• present a query recommendation, generated by a query recommendation model, comprising the query, to the user, the query recommendation model generated utilizing a machine learning method trained using a query-goal-mission structure;
• calculate a query recommendation user satisfaction metric associated with the feature and the query recommendation based upon user interaction of the user with the query recommendation;
• present a second query recommendation, generated by a second query recommendation model, comprising the query, to the user, the query recommendation model generated utilizing the machine learning method trained using the query-goal-mission structure;
• calculate a second query recommendation user satisfaction metric associated with the feature and the second query recommendation based upon second user interaction of the user with the second query recommendation;
• calculate a user satisfaction metric based upon at least one of a swipe rate through content associated with at least one of the goal, the second goal or the mission, a number of content items viewed associated with at least one of the goal, the second goal or the mission, a dismissal rate of the content associated with at least one of the goal, the second goal or the mission, a number of content items the user interacts associated with at least one of the goal, the second goal or the mission, or an average scroll rate through the content associated with at least one of the goal, the second goal or the mission;
• and
• based upon the user satisfaction metric, the query recommendation user satisfaction metric and the second query recommendation user satisfaction metric, select the query recommendation model that generated the query recommendation used to calculate the query recommendation user satisfaction metric associated with the feature and the query recommendation, but not the second query recommendation model that generated the second query recommendation used to calculate the second query recommendation user satisfaction metric associated with the feature and the second query recommendation, from amongst a plurality of query recommendation models for subsequent use for presenting query recommendations to users.

A non-transitory computer readable medium comprising instructions which when executed perform a method for improving mobile searches, comprising:

• identifying a feature, generated by a user within a time period on a client device, based upon an identification of the user;
• evaluating the feature to identify a goal associated with the feature;
• identifying a mission associated with the goal;
• identifying a second goal associated with the mission, the second goal having goal similarity within a similarity range of the goal;
• identifying a query associated with the second goal, the query having a weighted common goal similarity within a goal similarity range of the feature;
• evaluating a result, presented to the user, for the query, comprising elements ranked by a search result ranking model, the search result ranking model generated utilizing a machine learning method trained using a query-goal-mission structure;
• calculating a result user satisfaction metric associated with the result based upon user interaction of the user with the result;
• evaluating a second result, presented to the user, for the query, comprising second elements ranked by a second search result ranking model, the second search result ranking model generated utilizing the machine learning method trained using the query-goal-mission structure;
• calculating a second result user satisfaction metric associated with the second result based upon second user interaction of the user with the second result,
• the calculating a second result user satisfaction metric based upon at least one of a swipe rate through content associated with at least one of the goal, the second goal or the mission, a number of content items viewed associated with at least one of the goal, the second goal or the mission, a dismissal rate of the content associated with at least one of the goal, the second goal or the mission, a number of content items the user interacts associated with at least one of the goal, the second goal or the mission, or an average scroll rate through the content associated with at least one of the goal, the second goal or the mission;
• and
• responsive to the result user satisfaction metric exceeding the second result user satisfaction metric, selecting the search result ranking model that ranked the elements of the result used to calculate the result user satisfaction metric associated with the result, but not the second search result ranking model that ranked the second elements of the second result used to calculate the second result user satisfaction metric associated with the second result, from amongst a plurality of search result ranking models for subsequent use for presenting one or more results to users.