Intelligent data ranking system based on multi-facet intra and inter-data correlation and data pattern recognition
This disclosure is directed generally to an automatic intelligent electronic data processing system, platform, and method for computerized multi-facet data pattern recognition and ranking, and particularly to intelligently personalizing recommendation of data items for consumption by a particular entity based on past data consumption history of the entity and/or other entities via machine recognition of intra and/or inter-entity data item selection correlations. Such personalized recommendation may be based on a multi-facet ranking of the data items by integrating various intra-entity and inter-entity correlations and patterns in data item consumption into a quantifiable entity-specific ranking score for each data item that may potentially be selected for consumption by a particular entity.
1 . A system comprising:
a memory circuitry for storing computer instructions;
a network interface circuitry; and
a processor in communication with the network interface circuitry and the memory circuitry, the processor configured to execute the computer instructions to:
retrieve a plurality of lists, each of the plurality of lists being associated with a selection event and comprising a set of data items among a data item pool selected by one of a plurality of entities for the selection event;
generate, independent of identities of the plurality of entities, a cluster representation of the plurality of lists, based on a computerized pattern recognition algorithm;
extract a subset of the plurality of lists associated with a target entity to form a target list set;
automatically generate a ranking of one or more data items in the data item pool based on the cluster representation and the target list set;
automatically generate a recommendation among the one or more data items for the target entity based on the ranking; and
present, on a display device of a user equipment of the target entity, the subset of the plurality of lists with the generated recommendation for the target entity as a reduced presentation of a full menu.
2 . The system of claim 1 , wherein the cluster representation of the plurality of lists comprises a one or more list clusters, each of the one or more list clusters being associated with a subset of the plurality of lists.
3 . The system of claim 2 , wherein each of the one or more list clusters comprises a textual concatenation of data items included in the subset of the plurality of lists.
4 . The system of claim 3 , wherein the computerized pattern recognition algorithm is configured to cluster the plurality of lists based on inter-list similarities.
5 . The system of claim 4 , wherein the computerized pattern recognition algorithm comprises a K-mean computation.
6 . The system of claim 3 , wherein at least two subsets of the plurality of lists associated with different list clusters share a same list among the plurality of lists.
7 . The system of claim 6 , where to automatically generate a ranking of the one or more data items for the target entity comprises, for a current data item in the one or more data items, generate a data item ranking score associated with the target entity.
8 . The system of claim 7 , wherein the data item ranking score associated with the target entity for the current data item comprises a summation of cluster-wise sub-ranking scores for the target entity over each list cluster in a target list cluster set, the target list cluster set comprising list clusters among the one or more list clusters that include at least one list in the target list set.
9 . The system of claim 8 , wherein the cluster-wise sub-ranking score of the current data item for a current list cluster and associated with the target entity comprises a multiplication of at least:
a list count representing a number of lists associates with the target entity in the current list cluster;
a word count representing a number of times for at least one word in a textual form of the current data item appearing in the current list cluster; and
a data item count representing a number of times the current data item appearing in the target list set.
10 . The system of claim 9 , wherein the cluster-wise sub-ranking score of the current data item for a current list cluster and associated with the target entity is generated by additional multiplying a feedback score maintained for the current data item and the target entity, the feedback score indicating an extent to which the current data item was previously recommended to and selected by the target entity.
11 . The system of claim 10 , wherein the processor is configured to execute the computer instructions further to:
storing the recommendation under an entity identification of the target entity;
after providing the recommendation to the target entity, monitor a selection of data items from the data item pool by the target entity;
in response to a data item in listed in the recommendation being selected by the target entity, increment a feedback score associated with the data item and the target entity by a predetermined value.
12 . A method comprising:
retrieving a plurality of lists, each of the plurality of lists being associated with a selection event and comprising a set of data items among a data item pool selected by one of a plurality of entities for the selection event;
generating, independent of identities of the plurality of entities, a cluster representation of the plurality of lists, based on a computerized pattern recognition algorithm;
extracting a subset of the plurality of lists associated with a target entity to form a target list set;
automatically generating a ranking of one or more data items in the data item pool based on the cluster representation and the target list set;
automatically generating a recommendation among the one or more data items for the target entity based on the ranking; and
present, on a display device of a user equipment of the target entity, the subset of the plurality of lists with the generated recommendation for the target entity as a reduced presentation of a full menu.
13 . The method of claim 12 , wherein the cluster representation of the plurality of lists comprises a one or more list clusters, each of the one or more list clusters being associated with a subset of the plurality of lists.
14 . The method of claim 13 , wherein each of the one or more list clusters comprises a textual concatenation of data items included in the subset of the plurality of lists.
15 . The method of claim 14 , wherein the computerized pattern recognition algorithm is configured to cluster the plurality of lists based on inter-list similarities.
16 . The method of claim 13 , wherein at least two subsets of the plurality of lists associated with different list clusters share a same list among the plurality of lists.
17 . The method of claim 16 , where automatically generating a ranking of the one or more data items for the target entity comprises, for a current data item in the one or more data items, generating a data item ranking score associated with the target entity.
18 . The method of claim 17 , wherein the data item ranking score associated with the target entity for the current data item comprises a summation of cluster-wise sub-ranking scores for the target entity over each list cluster in a target list cluster set, the target list cluster set comprising list clusters among the one or more list clusters that include at least one list in the target list set.
19 . The method of claim 18 , wherein the cluster-wise sub-ranking score of the current data item for a current list cluster and associated with the target entity comprises a multiplication of at least:
a list count representing a number of lists associates with the target entity in the current list cluster;
a word count representing a number of times for at least one word in a textual form of the current data item appearing in the current list cluster; and
a data item count representing a number of times the current data item appearing in the target list set.
20 . A system comprising:
a memory circuitry for storing computer instructions;
a network interface circuitry; and
a processor in communication with the network interface circuitry and the memory circuitry, the processor configured to execute the computer instructions to:
retrieve a plurality of lists, each of the plurality of lists being associated with a selection event and comprising a set of data items among a data item pool selected by one of a plurality of entities for the selection event;
clustering the data item pool to generate a plurality of data item clusters based on a pattern recognition algorithm to identify inter-item correlations and similarities according to the plurality of lists; at least one data item in the data item pool appearing in more than one of the plurality of data item clusters;
receiving an input data item from a target entity among the plurality of entities;
automatically identifying a reduced set of data item clusters that includes the input data item;
executing a term-frequency-inverse-document-frequency (TFIDF) algorithm with respect to the input data item and the reduced set of data item clusters to identify a set of candidate data items within the reduced set of data item clusters having TFIDF scores higher than a predefined threshold score; and
generating, on a display device of a user equipment of the target entity, a recommendation including the set of candidate data items to the target entity in response to the target entity completing typing an input data item.