Skills ontology creation

Abstract

Disclosed in some examples are systems, methods, and machine readable mediums which allow for the automatic creation of a skills hierarchy. The skills hierarchy comprises an organization of a standardized list of skills into a hierarchy that describes category relationships between the skills in the hierarchy. The category relationships may include no relationships, parent relationships, and child relationships. A skill may be considered a parent of another skill if the parent skill describes a broader category of skill that includes the child. Other relationships such as grandparent (e.g., a parent's parent), great-grandparent, grandchild, great grandchild and so on may be defined inferentially as well. In some examples, the constructed hierarchy may be organized with broader skills at higher levels and narrower skills at lower levels.

1 . A method for creating a hierarchy of skills, the method comprising:
 training a skill classifier using training data and feature data, the training data comprising a training set of skill pairs, each skill in the training set of skill pairs selected from a predetermined list of skills, each particular skill pair in the training set of skill pairs labeled to indicate whether a parent-child category relationship exists between the skills in the particular skill pair in the training set of skill pairs, the feature data describing attributes about the skills in the training set of skill pairs, the attributes about the skills including data derived from member profiles of a social networking service;   determining for each particular skill pair in a target set of skill pairs, using the skill classifier and feature data for the skills in each particular skill pair, a probability that a parent-child category relationship exists in each particular skill pair, the target set of skill pairs including skills from the predetermined list of skills, the target set of skill pairs including at least one skill pair that is not in the training data;   collecting, in response to the determining the probability for each particular skill pair, a collection of probabilities for the target set of skill pairs; and   constructing a hierarchy of skills based upon the collection of probabilities for the target set of skill pairs.   
     
     
         2 . The method of  claim 1 , wherein the skill classifier is a random forest classifier. 
     
     
         3 . The method of  claim 1 , wherein the feature data describing attributes about the skills in the skill pairs comprises data describing a co-occurrence of each skill in the skill pairs in the member profiles of the social networking service. 
     
     
         4 . The method of  claim 3 , wherein the feature data includes a closeness centrality metric based upon a probability of co-occurrence between skills in the particular skill pair. 
     
     
         5 . The method of  claim 1 , comprising:
 labelling each particular skill pair of the training set of skill pairs based upon hierarchy information in a free text section of the member profiles of the social networking service.   
     
     
         6 . The method of  claim 1 , comprising:
 determining that a first probability for a first skill in the skill pair being the parent of a second skill in the skill pair is above a predetermined threshold and that a second probability for the second skill in the skill pair being the parent of the first skill in the skill pair is above the predetermined threshold, and in response:   assigning the first skill as the parent if the first probability is greater than the second probability and assigning the second skill as the parent if the second probability is greater than the first probability.   
     
     
         7 . The method of  claim 1 , comprising:
 presenting the hierarchy of skills on a member search screen of a social networking service;   receiving a selection of a skill in the hierarchy of skills; and   determining members of the social networking service that match search criteria, the search criteria comprising the selected skill; and   presenting the determined members of the social networking service that match the search criteria to a searcher.   
     
     
         8 . A non-transitory machine-readable medium for creating a hierarchy of skills, the machine-readable medium comprising instructions, which when performed by a machine, cause the machine to perform operations comprising:
 training a skill classifier using training data and feature data, the training data comprising a training set of skill pairs, each skill in the training set of skill pairs selected from a predetermined list of skills, each particular skill pair in the training set of skill pairs labeled to indicate whether a parent-child category relationship exists between the skills in the particular skill pair in the training set of skill pairs, the feature data describing attributes about the skills in the training set of skill pairs, the attributes about the skills including data derived from member profiles of a social networking service;   determining for each particular skill pair in a target set of skill pairs, using the skill classifier and feature data for the skills in each particular skill pair, a probability that a parent-child category relationship exists in each particular skill pair, the target set of skill pairs including skills from the predetermined list of skills, the target set of skill pairs including at least one skill pair that is not in the training data;   collecting, in response to the determining the probability for each particular skill pair, a collection of probabilities for the target set of skill pairs; and   constructing a hierarchy of skills based upon the collection of probabilities for the target set of skill pairs.   
     
     
         9 . The non-transitory machine-readable medium of  claim 8 , wherein the skill classifier is a random forest classifier. 
     
     
         10 . The non-transitory machine-readable medium of  claim 8 , wherein the feature data describing attributes about the skills in the skill pairs comprises data describing a co-occurrence of each skill in the skill pairs in the member profiles of the social networking service. 
     
     
         11 . The non-transitory machine-readable medium of  claim 10 , wherein the feature data includes a closeness centrality metric based upon a probability of co-occurrence between skills in the particular skill pair. 
     
     
         12 . The non-transitory machine-readable medium of  claim 8 , wherein the operations further comprise:
 labelling each particular skill pair of the training set of skill pairs based upon hierarchy information in a free text section of the member profiles of the social networking service.   
     
     
         13 . The non-transitory machine-readable medium of  claim 8 , wherein the operations further comprise:
 determining that a first probability for a first skill in the skill pair being the parent of a second skill in the skill pair is above a predetermined threshold and that a second probability for the second skill in the skill pair being the parent of the first skill in the skill pair is above the predetermined threshold, and in response:   assigning the first skill as the parent if the first probability is greater than the second probability and assigning the second skill as the parent if the second probability is greater than the first probability.   
     
     
         14 . The non-transitory machine-readable medium of  claim 8 , wherein the operations further comprise:
 presenting the hierarchy of skills on a member search screen of a social networking service;   receiving a selection of a skill in the hierarchy of skills; and   determining members of the social networking service that match search criteria, the search criteria comprising the selected skill; and   presenting the determined members of the social networking service that match the search criteria to a searcher.   
     
     
         15 . A system for creating a hierarchy of skills, the system comprising:
 a processor; and   a memory communicatively coupled to the processor, the memory comprising instructions, which when performed by the processor cause the system to perform operations of:   training a skill classifier using training data and feature data, the training data comprising a training set of skill pairs, each skill in the training set of skill pairs selected from a predetermined list of skills, each particular skill pair in the training set of skill pairs labeled to indicate whether a parent-child category relationship exists between the skills in the particular skill pair in the training set of skill pairs, the feature data describing attributes about the skills in the training set of skill pairs, the attributes about the skills including data derived from member profiles of a social networking service;   determining for each particular skill pair in a target set of skill pairs, using the skill classifier and feature data for the skills in each particular skill pair, a probability that a parent-child category relationship exists in each particular skill pair, the target set of skill pairs including skills from the predetermined list of skills, the target set of skill pairs including at least one skill pair that is not in the training data;   collecting, in response to the determining the probability for each particular skill pair, a collection of probabilities for the target set of skill pairs; and   constructing a hierarchy of skills based upon the collection of probabilities for the target set of skill pairs.   
     
     
         16 . The system of  claim 15 , wherein the skill classifier is a random forest classifier. 
     
     
         17 . The system of  claim 15 , wherein the feature data describing attributes about the skills in the skill pairs comprises data describing a co-occurrence of each skill in the skill pairs in the member profiles of the social networking service. 
     
     
         18 . The system of  claim 17 , wherein the feature data includes a closeness centrality metric based upon a probability of co-occurrence between skills in the particular skill pair. 
     
     
         19 . The system of  claim 15 , wherein the operations further comprise:
 labelling each particular skill pair of the training set of skill pairs based upon hierarchy information in a free text section of the member profiles of the social networking service.   
     
     
         20 . The system of  claim 15 , wherein the operations further comprise:
 determining that a first probability for a first skill in the skill pair being the parent of a second skill in the skill pair is above a predetermined threshold and that a second probability for the second skill in the skill pair being the parent of the first skill in the skill pair is above the predetermined threshold, and in response:   assigning the first skill as the parent if the first probability is greater than the second probability and assigning the second skill as the parent if the second probability is greater than the first probability.   
     
     
         21 . The system of  claim 15 , wherein the operations further comprise:
 presenting the hierarchy of skills on a member search screen of a social networking service;   receiving a selection of a skill in the hierarchy of skills; and   determining members of the social networking service that match search criteria, the search criteria comprising the selected skill; and   presenting the determined members of the social networking service that match the search criteria to a searcher.