System and method for context and sequence aware recommendation
The present disclosure provides a system and method for recommending context and sequence aware based training set to a user. The system identifies various items and keywords of a plurality of earlier trainings of the users' interest and generates a context and sequence aware recommendation model based on the context of the identified keywords. It uses a collapsed Gibbs Sampling as in generative modelling for prior trainings. Further, it applies the context and sequence aware recommendation model on various keywords that are of users' interest. The context and sequence aware recommendation model infers a plurality of subsequent trainings based on context derived from the keywords. In addition to this, the model is generated to rank the inferred plurality of subsequent topics using a probability distribution over subsequent keywords. At the last, it recommends at least one topic to the user based on ranking of the plurality of trainings.
1. A system for recommending at least one context and sequence aware training to at least one user of an enterprise, wherein the system comprising:
a processor;
a memory coupled to the processor, wherein the processor is capable of executing a plurality of modules stored in the memory;
an accessing module configured to access a plurality of historical training data, wherein the historical training data is of the user's interest comprising a plurality of items and one or more contextual keywords pertaining to trainings of interest specific to the user;
an identification module configured to identify one or more contextual keywords and one or more item bigrams from the plurality of items and the one or more contextual keywords pertaining to the trainings of interest specific to the user, wherein the one or more item bigrams are in a predefined sequence of two items associated with the plurality of items;
an analyzing module configured to identify one or more topic memberships of each user and one or more topic memberships for each item and each contextual keyword of the historical training data, wherein the topic membership is a probability distribution of one or more users over the plurality of historical training data;
a context and sequence aware recommendation model is generated in the form of the topic membership from the accessed plurality of historical training data by applying a collapsed Gibbs sampling technique over the plurality of items, one or more item bigrams and one or more context keywords, wherein the context and sequence aware recommendation model builds on a sequential topic model and adds the context to the sequential topic model, wherein the sequential topic model is implemented such that a sequence of preferred items by the user are generated using sampled topic from the user-specific distribution of topics and the previous item of the sequence; and
a recommendation module configured to apply the generated context and sequence aware recommendation model along with the one or more topic memberships of the user to recommend at least one subsequent training data based on at least one of the context keyword derived from the plurality of items and one or more item bigrams associated with the plurality of items of the plurality of historical training data, wherein the recommendation module optimizes the recommendations, using generated context and sequence aware recommendation model, by computing a score for every possible next item “s” and ranking the possible next item ‘s’ based on the score given by the last item of the sequence of preferred items or topics pertaining to topic memberships within the context, and given the sequence of preferred items or the topics “d”, the context “c” and the last preferred item “r”.
2. The system of claim 1 , wherein the plurality of historical training data includes one or more previous training data completed by at least one employee of the enterprise and one or more certification courses needed to an employee for a particular role in the enterprise.
3. The system of claim 1 , wherein the context and sequence aware recommendation model learns one or more topic memberships for each employee as well as for each item bigram and context keyword associated with each historical training data.
4. The system of claim 1 , wherein a user includes a new employee to the enterprise or an existing employee of the enterprise.
5. A method for recommending at least one context and sequence aware training to at least one user of an enterprise, wherein the method comprising:
accessing a plurality of historical training data, wherein the historical training data is of the users' interest comprising a plurality of items and one or more contextual keywords pertaining to trainings of interest specific to the user;
identifying one or more contextual keywords and one or more item bigrams from the plurality of items and the one or more contextual keywords pertaining to the trainings of interest specific to the user, wherein the one or more item bigrams are in a predefined sequence of two items associated with the plurality of items;
identifying one or more topic memberships of each user and one or more topic memberships for each item and each contextual keyword of the historical training data, wherein the topic membership is a probability distribution of one or more users over the plurality of historical training data;
generating a context and sequence aware recommendation model in the form of the topic membership from the accessed plurality of historical training data by applying a collapsed Gibbs sampling technique over the plurality of items, one or more item bigrams and one or more context keywords, wherein the context and sequence aware recommendation model builds on a sequential topic model and adds the context to the sequential topic model, wherein the sequential topic model is implemented such that a sequence of preferred items by the user are generated using sampled topic from the user-specific distribution of topics and the previous item of the sequence; and
applying the generated context and sequence aware recommendation model along with the one or more topic memberships of the user to recommend at least one subsequent training data based on at least one of the context keyword derived from the plurality of items and one or more item bigrams associated with the plurality of items of the plurality of historical training data, wherein the recommendation module optimizes the recommendations, using generated context and sequence aware recommendation model, by computing a score for every possible next item “s” and ranking the possible next item ‘s’ based on the score given by the last item of the sequence of preferred items or topics pertaining to topic memberships within the context, and given the sequence of preferred items or the topics “d”, the context “c” and the last preferred item “r”.
6. The method of claim 5 , wherein the plurality of historical training data includes one or more previous training data completed by at least one employee of the enterprise and one or more certification courses needed to an employee for a particular role in the enterprise.
7. The method of claim 5 , wherein the context and sequence aware recommendation model learns one or more topic memberships for each employee as well as for each item bigram and context keyword associated with each historical training data.
8. The method of claim 5 , wherein a user includes a new employee to the enterprise or an existing employee of the enterprise.
9. A non-transitory computer readable medium storing instructions for recommending at least one context and sequence aware training to at least one user of an enterprise, the instructions comprise:
accessing a plurality of historical training data, wherein the historical training data is of the users' interest comprising a plurality of items and one or more contextual keywords pertaining to trainings of interest specific to the user;
identifying one or more contextual keywords and one or more item bigrams from the plurality of items and the one or more contextual keywords pertaining to the trainings of interest specific to the user, wherein the one or more item bigrams are in a predefined sequence of two items associated with the plurality of items;
identifying one or more topic memberships of each user and one or more topic memberships for each item and each contextual keyword of the historical training data, wherein the topic membership is a probability distribution of one or more users over the plurality of historical training data;
generating a context and sequence aware recommendation model in the form of the topic membership from the accessed plurality of historical training data by applying a collapsed Gibbs sampling technique over the plurality of items, one or more item bigrams and one or more context keywords wherein the context and sequence aware recommendation model builds on a sequential topic model and adds the context to the sequential topic model, wherein the sequential topic model is implemented such that a sequence of preferred items by the user are generated using sampled topic from the user-specific distribution of topics and the previous item of the sequence; and
applying the generated the generated context and sequence aware recommendation model along with the one or more topic memberships of the user to recommend at least one subsequent training data based on at least one of the context keyword derived from the plurality of items and one or more item bigrams associated with the plurality of items of the plurality of historical training data, wherein the recommendation module optimizes the recommendations, using generated context and sequence aware recommendation model, by computing a score for every possible next item “s” and ranking the possible next item ‘s’ based on the score given by the last item of the sequence of preferred items or topics pertaining to topic memberships within the context, and given the sequence of preferred items or the topics “d”, the context “c” and the last preferred item “r”.