The application for bulk resume extraction and rating 

The purpose of this project is to build an Automated Resume Extraction, Bulk resume extraction and Classifier Service which will be built on Google's Cloud.Large enterprises and head-hunters receive several thousands of resumes from job applicants every day. HRs And Managers go through a hundreds of resumes manually. Resumes or Profiles are unstructured documents and have typically number of different formats (eg: .doc, .pdf).As a result manually reviewing multiple profiles is a very time consuming processes. How to ensure you have the Appropriate Candidate in the right jobs at the right time. Typically there is also a challenge of identifying fake profiles and filtering the profiles into a valid set of profiles that the company would like to talk to. This is a significant problem faced by large companies today in the market.

Manual extraction of information and allocating jobs without proper skill assessment. Costly extraction systems  which do keyword search with type and extraction limitations. Forcing candidates to fill templates and keep updating the templates as per job profiles.

The goal of this project is that the system will automatically do Bulk resume extraction. extract the resume content and store it  in a structure form within the Data Store Classification algorithms will be run on the profiles to identify profile categories or classes.Users can upload specific requirements and identify probable profile matches.

Introduction

Large enterprises and head-hunters receive several hundreds of resumes from job applicants every day. In general there is no standard format in which a resume can be written To induce standards so that the resumes can be electronically cataloged and searched, enterprises force job seekers to fill an online template. While this process helps the enterprise to effortlessly and quickly search for the right applicant, it induces unnecessary constraint on the applicant to fill in a different templates each time depending on the enterprise to which they are applying. A major problem associated with this approach is that the applicant is forced to tune their resume to match the style of the template which might not be able to capture all the details that the applicant might wish to display on their resume. Additionally, for the enterprise, the online template needs to be changed with time because of newer job descriptions or job types. Ideally, an enterprise would do away with forcing its applicants to fill in a predefined template provided they had access to a system that could extract the required information, both structured and unstructured, from any format of resume automatically. The benefit of such a system is that it would support automatic construction of an electronic resume database and would enable quick processing of resumes received by searching and routing resumes to appropriate destinations. Automatic extraction of information from resumes with high precision and recall is not an easy task essentially because of the non-standardization of resume structure. In spite of constituting a restricted domain, resumes can be written in multitude of formats (e.g. structured tables or plain texts) and in different file types (e.g. .tx t, .pdf, .doc(x) etc.).

Aims and Objective

The system aims to aid a large enterprise by removing the manual effort in screening resumes received by them to ascertain the suitability of candidate.We describe the system that is capable of automatically extracting relevant information from a resume and pushing the information into a database .The complete system is web enabled to make it reachable to a large number of people within the company.

This system will help us improve the performance of various companies in terms of precision and accuracy of the organization and will reduce the manual efforts taken by various employees and In-turn improve the Productivity of the organization.

In general there is no standard format in which a resume can be written To induce standards so that the resumes can be electronically cataloged and searched, enterprises force job seekers to fill an online template. While this process helps the enterprise to effortlessly and quickly search for the right applicant, it induces unnecessary constraint on the applicant to fill in a different templates each time depending on the enterprise to which they are applying. A major problem associated with this approach is that the applicant is forced to tune their resume to match the style of the template which might not be able to capture all the details that the applicant might wish to display on their resume. Additionally, for the enterprise, the online template needs to be changed with time because of newer job descriptions or job types. Resume Information Extraction is an application area of Information Extraction (IE) which is a process that automatically extracts predefined types of information from unstructured documents. Information Extraction also includes structuring, grouping and preparing found data to populate a database. Resume information extraction, also called resume parsing, enables extraction of relevant information from resumes which have relatively structured form. Although, there are many commercial products on resume information extraction,there has been surprisingly little published research work on this area.Some of the commercial products include Sovren Resume/CV Parser, Akken Staffing , ALEX Resume parsing , Resume-Grabber Suite and Daxtra CVX. There is a little information in specification of these products about methods and algorithms used for information extraction.

System Requirements

Processor: Pentium4 or Higher

RAM: 512 MB Hard Disk

Space: 50 MB

Software Requirements

Platform: platform independent

Programming Language: JAVA

Java Tool: Eclipse

Apache Tomcat

JDK

System Architecture

Figure 1: Architecture Diagram

The system is a web based client-server which is capable of automatically extracting information from resumes in English language and populating a structured database. The complete system consists of several modules as depicted in Fig. 1. It comes with an interface which allows for searching resumes populated in the database. The information Classification module is by and large the most significant component of the system.

Input module-The input module is a web interface which allows batch upload of resumes to the system. The system, has no constraint on the resume style or structure. The input module is additionally capable of accepting multiple resumes in the form of a .zip,.tgz, .7z, .Z, .gz file.

Information Extraction-The information extraction module is capable of extracting important relevant information from a free format resume automatically. The database build module populates the database with the extracted information and builds a resume database. Almost all resumes are unique in their structure and hence dissimilar, but one can assume a typical resume to have an overall hierarchical layered structure. The first layer is composed of several general information blocks such as personal information,education etc. The second layer of structure is within the first layer and contains specific information corresponding to the layer 1. For example, the layer 1 personal information block consists of layer 2 information like name, address and e-mail. While this might not be true for all the resumes, the structure seems to be retained in the bulk of resumes. Additionally, the location of the information (like name, age etc) in resumes vary Information extraction module is composed of several sub modules, each of which performs the task of extracting specific information. The main sub modules are (a) Qualification module, (b) Skill module (c) Experience module and (d) personal information extraction. While the qualification extraction sub-module extracts the graduating university name, degree and the class obtained. The skills extraction module extracts the skills of the candidate. Experience extraction module is capable or extracting the total experience, even when this information is not explicitly mentioned in the resume of the candidate. The name extraction module extracts applicant name and other information like date-ofbirth,email-id and passport number. The extraction process uses a set of language processing technique significantly from resume to resume. Our system can work on both layered structure and unstructured resumes.

The information extraction module is capable of extracting automatically from any given resume, information like, total experience, date-of-birth, passport number,email-id.skill set and qualification. The information extraction uses a bunch of natural language processing algorithms to extract relevant information from a free English resume. The information extracted by the information extraction module is populated into a database by the database build module.The current search module enables a user to search resumes with some particular criteria in the resume database. However a natural language interface to search resumes to enable searches like ”Show me all the resumes that have more than 3 years of java experience” would be an ideal interface to have.The search module gives an interface to query the system for a specific resume.

The user can query the resume database based on a combination of the following criteria

• age of candidate
• qualification
• software skills
• previous experience

All the resumes matching the criteria are displayed with a summary of the selected resume. Further a hyperlink allows the user to view the complete resume in its original form.

Implementation

Information extraction

Information extraction (IE) is a type of information retrieval whose goal is to automatically extract structured information from unstructured and/or semi-structured machine-readable documents. Resume or a Candidate Profile is typically unstructured data. We need to extract information and convert this into standard structured formats so that we can analyze or query on this data in an effective manner.

Document preprocessing

            First we convert the input resume a in different file types (.doc,.docs,.pdf) to .txt format.

We  need to maintain one dimension table for storing all the keywords that may appear in the input resumes. Then we have to travers thorough the txt file which is obtained after processing the input resume. So that we can find the keywords present in input resume in txt format and store them in database for that particular resume.

Search Profile

            For given search criteria for resume we check in database for the presence of given input criteria keywords in the all input resumes. Depending on that we need to generate the search result. That search result contains name of resume, matching percent.

Jelastic Cloud Platform

            Jelastic is the next generation of java hosting platform which can run and scale any Java Application with no code changes required.With Jelastic it is easy to create cloude environment. With Jelastic, you don't have to worry about code changes or lock-in — and you don't have to worry about managing servers or resources. Jelastic auto-scales, is easy to deploy, and runs any Java app.

We intend to use cloud platform to provide following features to our system.

• Multi-tenacity: enables sharing of resources and costs across a large pool of users
• Web-service: e.g.: LinkedIn.
• Scalability: and Elasticity via dynamic (”on-demand”) provisioning of resources on a fine-grained, self-service basis near real-time, without users having to engineer for peak loads.Cross company service: many company branches can use it.