SKILLSKAN | AI CV MATCHING
Semantic engine
for candidates pre-screening
Send us your documents, we will analyze the content and return you the professional tags and skills identified
SKILLSKAN | AI CV MATCHING
Semantic engine
for candidates pre-screening
Send us your documents, we will analyze the content and return you the professional tags and skills identified
Thanks to the web server connection, it is only necessary to upload via FTP protocol the candidate’s CV you would like to analyze or the content of the job offer.
The document content is inserted into the semantic engine, which springs into action, enclosing four different actions in just a few seconds:
- data parsing and extraction;
- data normalization;
- semantic analysis;
- indexing.
Thanks to the web server connection, it is only necessary to upload via FTP protocol the candidate’s CV you would like to analyze or the content of the job offer.
The document content is inserted into the semantic engine, which springs into action, enclosing four different actions in just a few seconds:
- data parsing and extraction;
- data normalization;
- semantic analysis;
- indexing.
Data extraction and content categorization
Once a document is entered, typically a CV (whether it is a reading file, an image, or a scan), the system returns that document in text format (plain text or html).
In this first step the images are analyzed, thanks to an integrated software of “image processing” based on neural networks, which performs the CV parsing and makes a classification of the text content in sections, such as:
- Image of the candidate (if any)
- Candidate CV – with censored personal data
- Part of the CV referring to: experience, education, personal data, languages and other skills/hobbies/hobby
Data extraction and content categorization
Once a document is entered, typically a CV (whether it is a reading file, an image, or a scan), the system returns that document in text format (plain text or html).
In this first step the images are analyzed, thanks to an integrated software of “image processing” based on neural networks, which performs the CV parsing and makes a classification of the text content in sections, such as:
- Image of the candidate (if any)
- Candidate CV – with censored personal data
- Part of the CV referring to: experience, education, personal data, languages and other skills/hobbies/hobby
Data normalization and semantic analysis
The data are normalized in order to improve understanding and a grammar software conducts a logical analysis on the extracted content to identify the semantic meaning of each sentence in the text.
Our calculation algorithm searches and indexes professional skills (the tasks that uniquely identify a profession) in the analyzed text.
- Given a text (plain text or html) a title and a language, the system will provide a set of semantic tags linked to the text and its weight.
- Given a text (plain text or html) or a keyword and a language, the system will provide a clean text to optimize fulltext searches.
Data normalization and semantic analysis
The data are normalized in order to improve understanding and a grammar software conducts a logical analysis on the extracted content to identify the semantic meaning of each sentence in the text.
Our calculation algorithm searches and indexes professional skills (the tasks that uniquely identify a profession) in the analyzed text.
- Given a text (plain text or html) a title and a language, the system will provide a set of semantic tags linked to the text and its weight.
- Given a text (plain text or html) or a keyword and a language, the system will provide a clean text to optimize fulltext searches.
Indexing and matching
Professional skills are analyzed and evaluated by calculation algorithms, to then be accepted or rejected based on parameters of professional proximity.
Currently the semantic analysis system is available in 6 languages: Italian, English, French, German, Spanish, Portuguese.
The algorithm runs an extraction of the professional tags contained in the job offers; then it compares them with the skills found in the candidates’ CVs and returns the matching of suitable CVs for the selected talent pool.
Indexing and matching
Professional skills are analyzed and evaluated by calculation algorithms, to then be accepted or rejected based on parameters of professional proximity.
Currently the semantic analysis system is available in 6 languages: Italian, English, French, German, Spanish, Portuguese.
The algorithm runs an extraction of the professional tags contained in the job offers; then it compares them with the skills found in the candidates’ CVs and returns the matching of suitable CVs for the selected talent pool.