Method and apparatus for talent-post matching and computer readable storage medium
Embodiments of the disclosure provide a method and an apparatus for talent-post matching, a device and a medium, which relates to the field of information matching and recruitment. The method includes: determining a resume feature representing a resume to be predicted according to a post submitted and the resume to be predicted; and predicting a matching degree between the resume to be predicted and the post by using a pre-trained predictive model according to a post feature of the post and the resume feature. With the method and the apparatus for talent-post matching, the device and the medium provided in embodiments of the present disclosure, the resume and the post may be matched automatically.
1. A method for talent-post matching, comprising:
determining a first resume feature on a word level according to a resume to be predicted;
determining a first post feature on a word level according to a post;
determining a second post feature by associating features in the first post feature with a plurality of first weights;
determining a post feature by adding up a plurality of second post features of the post set with a plurality of second weights;
setting a plurality of third weights for features in the first resume feature according to a matching degree of the features in the first resume feature and the second post feature, associating the features in the first resume feature with the plurality of third weights, and determining the associated features as a second resume feature, in which the second resume feature represents a work experience in the resume to be predicted;
determining a resume feature representing all work experiences in the resume to be predicted according to the second resume feature and the post feature; and
predicting, according to the post feature of the post and the resume feature, a matching degree between the resume to be predicted and the post by using a pre-trained predictive model which is trained according to training samples, the training samples comprising posts and resumes to be trained.
2. The method according to claim 1 , wherein after predicting, according to the post feature of the post and the resume feature, the matching degree between the resume to be predicted and the post by using the pre-trained predictive model, the method further comprises:
performing a cause analysis on a predicted result of said predicted matching degree.
3. The method according to claim 2 , wherein performing the cause analysis on the predicted result of said predicted matching degree comprises at least one of:
distinctively displaying, in a text of the post, a word or a sentence corresponding to a sub-feature in the post feature according to the matching degree, the matching degree of the sub-feature in the post feature satisfying a preset post matching condition; and
distinctively displaying, in a text of the resume to be predicted, a word or a sentence corresponding to a sub-feature in the resume feature according to the matching degree, the matching degree of the sub-feature in the resume feature satisfying a preset resume matching condition.
4. An apparatus for talent-post matching, comprising:
one or more processors;
a storage device for storing one or more programs;
wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to:
determine a first resume feature on a word level according to a resume to be predicted;
determine a first post feature on a word level according to a post;
determine a second post feature by associating features in the first post feature with a plurality of first weights;
determine a post feature by adding up a plurality of second post features of the post set with a plurality of second weights;
set a plurality of third weights for features in the first resume feature according to a matching degree of the features in the first resume feature and the second post feature, associate the features in the first resume feature with the plurality of third weights, and determine the associated features as a second resume feature, in which the second resume feature represents a work experience in the resume to be predicted;
determine a resume feature representing all work experiences in the resume to be predicted according to the second resume feature and the post feature; and
predict, according to the post feature of the post and the resume feature, a matching degree between the resume to be predicted and the post by using a pre-trained predictive model which is trained according to training samples, the training samples comprising posts and resumes to be trained.
5. The apparatus according to claim 4 , wherein after the one or more processors predict, according to the post feature of the post and the resume feature, the matching degree between the resume to be predicted and the post by using the pre-trained predictive model, the one or more processors are configured to:
perform a cause analysis on a predicted result of said predicted matching degree.
6. The apparatus according to claim 5 , wherein the one or more processors perform the cause analysis on the predicted result of said predicted matching degree by performing at least one of:
distinctively displaying, in a text of the post, a word or a sentence corresponding to a sub-feature in the post feature according to the matching degree, the matching degree of the sub-feature in the post feature satisfying a preset post matching condition; and/or,
distinctively displaying, in a text of the resume to be predicted, a word or a sentence corresponding to a sub-feature in the resume feature according to the matching degree, the matching degree of the sub-feature in the resume feature satisfying a preset resume matching condition.
7. A non-transitory computer readable storage medium having stored thereon a computer program, wherein when the computer program is executed by a processor, the processor is cause to:
determine a first resume feature on a word level according to a resume to be predicted;
determine a first post feature on a word level according to a post;
determine a second post feature by associating features in the first post feature set with a plurality of first weights;
determine a post feature by adding up a plurality of second post features of the post set with a plurality of second weights;
set a plurality of third weights for features in the first resume feature according to a matching degree of the features in the first resume feature and the second post feature, associate the features in the first resume feature with the plurality of third weights, and determine the associated features as a second resume feature, in which the second resume feature represents a work experience in the resume to be predicted;
determine a resume feature representing all work experiences in the resume to be predicted according to the second resume feature and the post feature; and
predict, according to the post feature of the post and the resume feature, a matching degree between the resume to be predicted and the post by using a pre-trained predictive model which is trained according to training samples, the training samples comprising posts and resumes to be trained.