CA2388250C - Method for providing access to online employment information - Google Patents

Abstract

The present invention provides a method of managing employment data so as to provide access to the employment data via the Internet (18). The method including the steps of determining whether a web site (22, 24) contains employment data, formatting, parsing and storing the employment data and corresponding URL into a database, automatically searching the database (16) for matching employment data, and contacting the employer representative as to the matched employment data.

Description

METHOD FOR PROVIDING ACCESS TO
ONLINE EMPLOYMENT INFORMATION
FIELD OF INVENTION
The present invention relates to employment services and, in particular, to online recruiting or employment services.
BACKGROUND OF THE INVENTION
The rapid expansion of job postings on the Internet has created a large amount of employment related information, which spans hundreds of thousands of web sites.
Initially, companies began posting their open job positions on their own corporate web sites. A job seeker could then readily access new employment opportunities by visiting a company's web site. As an increasing number of company web sites began to post their open jobs, however, the job search process grew proportionally. For example, a job seeker searching for a "software developer" position would have had to identify and visit the web site of every company that might have such open job positions. Thus, this growth resulted in a task that was cumbersome and time consuming for the job seeker.
In order to help address these issues, job board web sites have evolved on the Internet. The original purpose of a job board was to provide a single web site where companies could visit to post their open job positions and job seekers could visit to search for new employment opportunities. The job board concept helped the job seekers by creating a central location that a job seeker could visit to search for jobs.
Unfortunately, however, the concept increased the work and cost for companies.
In addition to maintaining job postings on their own corporate web sites, companies were now required to visit the job board sites to repost, update and delete their job position information as appropriate. The accuracy of the job board information was affected when companies changed their job information, filled open position, etc., but failed to update the corresponding job board postings. These job boards also often charged a fee to the SUBSTITUTE SHEET (RULE 26) companies for this posting service. In addition, these job boards only contained job positions from companies that had actively posted jobs on the sites. In other words, companies that did not know about the job boards would have been prevented from listing the company's open positions and, consequently, eliminated opportunities for the job seekers as well as the company itself.
Most recently, the aggregation, accuracy, and freshness of job board postings have been addressed through various web spidering or crawling technologies. The technology of web site spidering or crawling consists of a process in which content from a set of source web sites is retrieved automatically. This content is typically retrieved for purpose of being indexed into a search engine web site in order to provide Internet users a central web site to use as a search tool. The type of content that is spidered is generally not filtered so the search engine web site often has indexed content from a wide variety of source web sites. New web sites that contain content to be spidered have to register with the search engine web site before their content is retrieved and indexed into the search engine. Once a new site is registered into the set of source web sites to spider, the search engine web site will periodically spider the site to search for new or updated content to index.
In these updated models, the job board periodically sends out spiders to the web sites of companies that register with the job board web site. The purpose of these spiders is to retrieve and input the latest job posting infonnation from the company web sites and thereby automatically update the job information listed on the job board. The method, however, creates a disadvantage for companies and job seekers because the sites do not post the numerous job positions from the companies that do not register with or know of the job board web site. As such, the Internet contains a vast amount of job postings which exist only on company job boards and which are not being collected and displayed by the job board web sites.
Another new approach to job posting aggregation is the master search engine site.
In this approach, the master web site collects a job seeker's search criteria and submits it to multiple other job board web sites. The master search engine site aggregates the individual sites and presents the results to the job seeker in a single format. An advantage to this method is that the job seeker only needs to visit a single site to perform a job search. The SUBSTITUTE SHEET (RULE 26) disadvantages of this approach are that, as described above, only a subset of the job board sites on the Internet are actually searched and individual company job postings are completely omitted. Furthermore, in these types of searches, the formatting of the results can vary thereby causing the job seeker to become confused when presented with search results.
An additional feature of prior art job board web sites is the electronic notification of new job opportunities. When a new job is posted that fits within his selected category infonnation, the job seeker automatically receives notification of the new job via email. A
limitation to this system is that user may miss employment opportunities which are filtered outside of the selected category information.
Another drawback of the prior art systems relate to the search engines used for identifying a position of interest to the job seeker. The prior art systems use a table, key word or boolean driven search engine. The search engines use a pull-down menu, keyword or boolean search methodology that has a limited ability to implement intelligent searches.
For instance, a job seeker may be in search of a position in a specific technical field. A
search of job postings with one or two keywords may identify many unrelated jobs. It may be very time consuming for the job seeker to review every identified job posting. The effort becomes even greater when compounded by the number of such searches to be completed at each of the numerous online employment sites. The job seeker may use additional keywords to reduce the number of unrelated job postings. However, the additional keywords often have the effect of reducing certain of the job postings, which may be of interest to the job seeker, but do not necessarily contain all of the designated keywords. In other words, the search strategy may have become too restrictive.
Therefore, the job seeker ends up accessing only a small fraction of jobs currently available on the Internet.
Along with the evolution of job board related web sites, the prior art systems have provided job seekers the ability to post electronically their resumes. These systems have increased the amount of resumes available online. This increase has created web sites, which collect resumes into searchable databases. These web sites often sell subscription access to their databases, which employers and recruiters purchase in order to search for qualified candidates. However, these web sites suffer from the same disadvantages and SUBSTITUTE SHEET (RULE 26) limitations as described in the job posting process: a) companies and job seekers must visit .
the web sites to add and update information; b) searches are limited to narrowly targeted keywords; and c) job seeker resumes are sorted into restrictive categories.
Furtheiniore, if these companies do not post at the job board web sites, without adequate traffic to their corporate web site and employment pages, employers cannot, on their own, reach a sufficient number of qualified candidates. As a result, the employers must choose to either pay the third party job board web sites to post a portion of their jobs online, making these opportunities accessible to a larger candidate pool, or miss many qualified candidate. Despite this investment, however, the factors listed above still limit the effectiveness of the job boards and prevent many qualified candidates from matching with the opportunities employers have paid to list.
In summary, there are deficiencies in the current state of the art in the Internet based employment process. The gap between job board listings and actual online jobs is growing rapidly. Companies develop and add recruiting pages to their own web sites much faster than the rate at which the top job boards add clients. Moreover, the gap between unique job board listings and unique jobs available online is expanding at an even faster pace, as companies that use job boards often post the same opening to between six and ten sites. Furthermore, the current web site job boards fail to aggregate completely all job postings on the Internet. Even the sites that aggregate a larger amount of the available job listings are limited by the search engine technology currently used by those job boards. In addition, the current prior art systems are deficient in their information exchange capabilities. Job board web sites rely on companies and/or job seekers to continually visit the job board web sites and update the applicable information.
SUMMARY OF THE INVENTION
The object of the present invention is a method of managing employment data to provide enhanced access via the Internet to the employment data.
A further object of the present invention is to provide a more thorough and precise searching of the employment data.
Still a further object of the present invention is to update automatically the employment data collected by the present invention.

SUBSTITUTE SHEET (RULE 26) Still yet a further object of the present invention is to format the employment data so as to allow for a more accurate and efficient search of the employment data.
Still yet a further object of the present invention is to match automatically users to fulfill employment needs.
In general, the present invention consists of several key subsystems. These subsystems are based on existing software technology, information spidering and concept based searching, which is new in its application to the Internet related employment industry.
The present invention builds on the technology of job spidering and aggregation and incorporates it into the employment field. For example, the working set of web sites which this system spiders includes the entire Internet directory ("Dot Corn database").
Thus, both companies and job boards are included in the job posting collection.
Furtheimore, the use of spidering technology is extended to resume collection as well as spidering of job postings. This allows the creation of a much more comprehensive and complete database of the available employment data.
The present invention also applies a concept based search engine to the employment search and match problem. As noted above, prior art search engine web sites are commonly based on keyword search engine technology. In its simplest form, a keyword search takes a set of comma delimited user input words and scans its document set for one or more word or partial word matches. Keyword searches, however, have been enhanced to include word count statistics, i.e., how often a word appears in a document increases its relevancy, and boolean operators, i.e., a user can search for specific terms to return documents that must contain both words. Unfortunately, these searches remain as simple word pattern matching technology, and the casual Internet user does not necessarily possess a clear understanding of query word relevancy or boolean logic.
In order to improve the user search experience, concept based search engines were created. The premise of a concept based search engine is that it is able to "learn" thematic information regarding the documents that it indexes. This learning is typically accomplished by applying Bayesian reasoning and neural network technology to each document when it is indexed. Users are often able to search the database by using full sentence, natural language queries instead of keyword sets and boolean logic.
As a concept SUBSTITUTE SHEET (RULE 26) based search engine learns its document set, it can also make distinctions and relations.
This learned infounation allows a user to search effectively for infonnation without knowing exactly what is being sought or how the query should be phrased.
Another important feature of a concept based search engine is that the user will always be provided with some form of results. The results from such a search engine are typically returned in descending weight order. A result with 100% weight is highly relevant to the user's query, while a result with 1% weight contains little or no relevance to the search. This behavior is a key feature of the concept based search engine, because it allows a programmatic decision to be made based on the "goodness" of a particular result.
The use of a concept based search engine in the present invention eliminates the need for the user to categorize a job posting or resume into a fixed category list and to rely on simple keyword based searches to find information, thereby providing an accurate and thorough search result. The present invention then automatically spiders job and resume related web sites for content, indexes the content into its concept based search engines, matches the content between jobs and resumes, and notifies companies and job seekers of new mutual opportunities. This process occurs continuously to maximize the timeliness and freshness of the information exchange.
Also, the present invention is able to accept a wide range of job posting formats and resume formats. The format of a job posting or resume will vary, often significantly, from web site to web site and job seeker to job seeker. By enhancing the process with newly developed software, which targets the online employment information, the system is able to index this diverse data into a common format. Once in a common format, matches within the data between job postings and resumes are efficiently performed.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a functional block diagram of the system of the present invention.
Fig. 2 shows a functional flowchart for creating and accessing a database of employment data available on the Internet.
Fig. 3 shows a flow chart for determining if the visited web sites meet the employment criteria.

SUBSTITUTE SHEET (RULE 26) Fig. 4 shows a flow chart for updating automatically the employment data stored in the database.
Fig. 5 shows a flow chart for formatting and parsing the employment data.
Fig. 6 shows a flow chart for adjusting the revisitation period of the visited web sites.
Fig. 7 shows a flowchart showing the aging and deletion step.
Fig. 8 shows a flow chart for collecting subscriber search criteria and conducting a concept-based search using the criteria.
Fig. 9 shows a flowchart of matching the employment data and notifying the users.
Fig. 10 provides a table depicting employment data.
DETAILED DESCRIPTION OF THE INVENTION
With reference to FIG 1, a system 10 of managing employment data is shown. The system 10 includes a dedicated spidering server 12, a dedicated search, retrieve and process server 14 and a database 16. The system 10 provides users (not shown) with the ability to search, via the Internet 18, for employment data located at public job boards 20, corporate web sites 22 and other web sites 24. Users are provided access to the system 10 via user Internet connections 26. The Internet connections 26 may be personal computers, for example.
The dedicated spidering server 12 is used to search the Internet for the employment data. FIG 10 provides a table showing an example of employment data 28 or information available via the Internet 18. Once the employment data is located, relevant information is loaded into the database 16. The dedicated search, retrieve and process server 14 provides the user the ability to search the database 16 for employment data. Users include corporation representatives seeking to fill a position, agents working for the corporations, as well as individuals seeking an employment position. The dedicated search, retrieve and process server 14 also conducts automatic searches of the database for matching employment data (i.e., matching jobs and resumes).
It will become clear from FIG 2 that the database 16 of FIG 1 represents multiple databases having individual functions. FIG. 2 discloses a process or functional block diagram of the present invention. In particular, FIG 2 discloses a process which dynamically retrieves and indexes large amounts of web employment data and processes this information in an efficient and timely manner. The Dot Corn database 30 contains a listing of all the active domain names on the Internet 18. The prequalify dictionary 32 consists of a concept based search engine that has been loaded with template documents to identify web pages that contain job posting or resume information. The site prequalification step 34 receives input from the Dot Corn database 30 and the prequalify dictionary 32. The site prequalification step 34 filters web sites that contain job postings or resumes. The output of step 34 includes URL records, which are stored in the active spider's database 36. Step 34 is shown in greater detail in FIG. 3. Begin site prequalification at step 3.1 Step 3.2 of FIG 3 begins with reading the prequalify dictionary 32. Step 3.3 reads the next record from the Dot Corn data base 30. Determine if end of records at step 3.4. If end, then done at step 3.14. Step 3.5 consists of determining whether the record is scheduled for a check. At step 3.6, each record is checked against the Internet domain named service (DNS) to verify whether an active web site exists for the domain name. Determine if DNS valid at step 3.7. In the event it is determine that an active web site does not exist, then step 3.13 consists of scheduling the web site or record for a future check. In the event the web site is active, step 3.8 consists of fetching the content of the web site. Step 3.10 consists of checking the site content against the prequalify dictionary 32. The prequalify dictionary 32 contains a concept base search engine which has been configured with template sample documents of job postings and resumes. Each page of site content that is retrieved at step 3.8 is presented as a query input to the prequalify dictionary concept based search engine at step 3.10.
Determine if check successful at step 3.11 and if so, add site to active spider database at step 3.12. The search engine returns a rated percent result, which indicates how relevant a particular site page is with respect to job postings or resumes. If a web site is determined to contain documents of sufficient relevancy, the site is stored in the active spider's database 36, enabling the site to be regularly spidered for its content. The retrieve content is stored in the spidered content database 38. If a web site does not exist or has no relevant content, it is scheduled at step 3.13 for a future check, at which time the site prequalification step 34 will revisit the site to repeat the foregoing process.
The site prequalification step 34 contains several key operating parameters, including the maximum number of pages to retrieve from a single web site, the amount of time to spend spidering a single web site and a threshold relevancy wait that is used to indicate whether the site contains job postings or resumes of related content.
Critical to this step is the configuration of the prequalify dictionary 32, as its document set is the mechanism that controls which web sites are accepted as valid and which are rejected. The architecture of a site group prequalification step 34 is readily scalable, as in practice several services can be operating in parallel on the Dot Corn data base 30 to perform the web site validation process. By scaling services in this manner, the information scan rate of the millions of records of the Dot Corn database 30 is easily controlled.
The periodic spidering step 40 of FIG 2 is responsible for running each of the spiders in the active spider's database 36 on a regular, scheduled basis. FIG
4 discloses the periodic spidering step 40 in greater detail. Begin periodic spidering at step 4.1. Step 4.2 consists of reading the next record from the active spider's database 36.
Determine if end of records at step 4.3 and if so, done at step 4.11. Step 4.4 determines whether the web site corresponding to the record is scheduled to be spidered. In the event the web site is scheduled to be spidered, step 4.5 fetches the site content. Step 4.7 determines whether the newly fetched content has changed from the corresponding content previously stored in the spidered content database 38 (FIG. 2) to determine whether the web site has changed.
If a change has occurred, the new content is stored in the spider content database 38 for further processing. If content changed, then update spidered content database at step 4.8.
If it is determined at step 4.6 that the spider fails when accessing a particular web site, step 4.9 consists of identifying the site as "failed" and removing the sit& from the active spider's database 36. Step 4.10 updates the Dot Corn database 30 to schedule the site to be requalified at a later time.
Step 40 is designed to run continuously to ensure that when the content of each source site changes, it is quickly updated in the spider content database 38.
Thus, the timeliness and freshness of the information is preserved. Step 40 is readily scalable, as in practice several services can be operated and parallel to perform this spidering process. As additional spiders are created, additional service can be added to handle the new load.
The content processing step 42 of FIG. 2 consists of further processing the content, which is temporarily stored in the spider content database 38. The processing dictionary 44 consists of a concept based search engine, which is similar to the prequalify dictionary 32.
The search engine has been loaded with additional template documents that enable spidered content to be parsed and scrubbed prior to being loaded into the searchable content database 46. The content processing step 42 is shown in greater detail in FIG 5.
The content processing step 42 is responsible for processing each retrieved document into a format that is suitable for indexing into the searchable content database 36. The processing dictionary 44 contains a concept based search engine, which has been configured with documents that contain specific job titles, job descriptions and resume descriptions. The dictionary 44 is used to measure the relevance of each spidered content document to determine whether it should be classified as a job-posting, resume or irrelevant, at which time the content is discarded. Another task of step 42 is the parsing and analysis of web pages, which contain multiple sets of information. For example, a single web page, which contains 15 different job postings, is broken down into 15 separate documents utilizing available advanced document parsing technology. Each document would contain its own title and specific job location information. The improved content results in a search experience that is clear and concise to the user.
Begin content processing at step 5.1. Step 5.2 consists of reading the processing dictionary 44. Step 5.3 consists of reading the next record from the spidered content data base 38. Determine if end of records at step 5.4 and if so, done at step 5.14.
Step 5.5 strips the document of its hypertext markup language (HTML) commands. The stripped document is evaluated by step 5.6 for its length requirements, and is scanned at step 5.7 and 5.8 to identify the location information (city, state, and zip code), and the e-mail address information.
The document is then presented as query input through the processing dictionary 44. The concept based search engine is used to further identify the document as a job posting or resume as well as determine its title information and amount of different information which the document may contain (see step 5.9). Documents that do not meet minimum relevancy requirements as a job posting or resume are discarded (step 5.10 and 5.12). Scan for document body content at step 5.11. Documents that pass the noted criteria are indexed into the searchable content database 46 as a job posting or resume (step 5.13).
After a document passes through this process, its record in the searchable content database 46 represents a uniform entry, which is consistent with the other records. The content processing step 42 is designed to run continuously as new information is placed into the spidered content database 38. Thus, the timeliness and freshness of the information is preserved. Step 42 is readily scalable, as in practice several servers can be operating in parallel to perform the content processing. As the input spidering process information flow increases, additional servers can be added to handle the new content processing load.
The spider adaptation step 48 of FIG 2 is responsible for dynamically adjusting the operating parameters of each spider. The adaptation step 48 is shown in greater detail in FIG. 6. Begin spider adaptation at step 6.1. Step 6.2 consists of reading the next site of which the content was previously processed and stored in the searchable content database 46. Determine whether end of sites at step 6.3 and if so, done at step 6.9 In the event it is determined at step 6.4 that the particular spider failed or retrieved irrelevant content (not job posting or resume related content), then step 6.10 sets the spider status as "failed" in the active spider data base 36, and at step 6.11, the Dot Corn data base 30 is updated to requalify the failed site at a later time.
Step 6.5 compares the content retrieved at step 6.2 with the content previously stored in the searchable content database 46. Step 6.6 determines whether the changed limit has been exceeded. Based on the amount of changes that have occurred, the spider schedule will be adjusted accordingly. In the event the change limit has been exceeded, then step 6.12 will set the spider to run again the following day. In the event the change limit has not exceeded, then step 6.7 and 6.8 will increase the spider frequency for that particular site by an additional day if the delay is presently less than 30 days. The spider adaptation step 48 is designed to run continuously as a feedback loop between the content processing step 42 and the periodic spidering step 40. Step 48 is readily scalable, as in practice several servers can be operating in parallel to perform this step 48.
As the input spidering process information flow increases, additional service can be added to handle the new load.
The aging and deletion step 50 is responsible for expiring old information in the searchable content database 46. The aging and deletion step 50 is shown in greater detail in FIG 7. Begin aging and deletion at step 7.1. Step 7.2 reads the next record from the searchable content data base 46. Determine if end of records at step 7.3 and if so, done at step 7.6. Step 7.4 determines whether the document date has expired. In the event the document date has expired, step 7.5 deletes the document from the searchable content database 46. Step 50 ensures that old web sites that have been removed from the Internet are identified, and their content document sets are purged from the overall system. The aging and deletion step 50 is designed to run continuously, and it is readily scalable, as in practice several servers can be operating in parallel to perform this aging and deletion step.
The result of the foregoing provides a searchable content database 46 of job positions and resumes, which may be "manually" searched by users as well as searched via an automatic process.

The "manual" search is initiated at the user search step 52 and continues with the concept phase step 54, the keyword phase step 56 and concludes with the search results 58. FIG 8 discloses additional details as to the user search. Begin user search at step 8.1.
Step 8.2 consists of reading the user search input. Step 8.3 determines whether the title, description or key words have been entered. If yes, determine if location entered at step 8.4. If location not entered, clear search start index at step 8.5. However, the user may further include information such as the city, state, range of location and number of results returned, etc. The concept phase step 54 occurs at step 8.6 whereupon concept searching is conducted upon the searchable content database 46 using the user input.
Determine if results at step 8.7. The results are processed at step 8.8 whereupon traditional text processes and techniques are used on the result to produce a filtered result set. Step 8.9 determines whether the quantity of the results meets the users specified quantity in order to determine whether the search may be concluded. If yes, then done at step 8.10. If the location is entered at step 8.4, then validate city and state location, determine if location valid, determine if location range entered, and if yes, build concept location search spec.
The user search step provides a front-end, manual interface for job seekers and employers or recruiters to search for employment data, i.e., job postings or resumes, respectively. The job seeker's search is provided as a free service, whereas the resume search is sold as a subscription service.
The user search is designed to run on user demand, and is readily scalable, as in practice several servers can be operating in parallel to service multiple user search requests. As the number of new users searching the system increases, additional servers can be added to handle the new load.
The automatic match step 60 is responsible for identifying matches between the employer's (job postings) and job seekers (resumes). As matches are identified, both the employer and job seeker are notified via e-mail. FIG 9 discloses the automatic match step 60 in greater detail.
Begin automatic match at step 9.1. Step 9.2 consists of reading the next new job posting from the searchable content database 46. Determine if end of records at step 9.3.
Step 9.4 consists of using the contents of the new job posting as query input to perform a concept based search on the resumes in the searchable content data base 46.
Determine if results found at step 9.5. The results of this search consist of a set of resumes that meet a relevant percent rate with respect to the job posting content. The candidates of these resumes are identified as "good matches" for a particular job posting. At steps 9.6 and 9.7, the employer corresponding to the new job posting and the candidates corresponding to the identified resumes, are contacted via e-mail.
Step 9.8 consists of reading the next new resume from the searchable content data base 46. Determine if end of records at step 9.1., and if so, done at step 9.14. At step 9.10, the contents of the new resume are used as query input to perform a concept based search on the job postings in the searchable content database 46. Determine if results found at step 9.11. The results of this search consist of a set of job postings that meet a relevant percent rate with respect to the resume content. The jobs are identified as "good matches"
for the particular candidate. Steps 9.12 and 9.13 consist of contacting the employers corresponding to the job posting results, and the candidate corresponding to the new resume.
When a candidate receives an e-mail message containing the job description(s), the candidate is able to access the job posting details, company information, etc.
free of charge. Once the candidate reviews this information, the candidate may choose to apply to a job, also free of charge. When an employer or recruiter receives the e-mail message identifying an eligible candidate(s) and the qualification summaries, the employer or recruiter may elect to purchase a web site subscription, which allows access to each candidate's resume and contact information. Furthermore, when an employer or recruiter subscribes to the web site and accesses various candidate information, the employer or recruiter may also elect to engage recruiting services to assist in pursuing the candidate.
The automatic match step 60 is designed to run continuously as new job postings and resumes are added to the searchable content database 46. The match step 60 is scalable, as in practice several servers can be operated in parallel to perform this matching and e-mail notification process. As the input information flow to the searchable content database 46 increases, additional servers can be added to handle the new load.

Claims (7)

1. A method of managing employment data so as to provide access to the employment data via the Internet, the method comprising:
in a system linked to the Internet, the system comprising at least a spidering server, a search, retrieve and process server, and a database, collecting the employment data from information available from a plurality of web sites on the Internet, each of the web sites represented by a unique Uniform Resource Locator (URL), the collecting comprising checking site content of each of the plurality of web sites against a prequalify dictionary containing a concept base search engine that is configured with template sample documents of job postings and resumes;
in the system, formatting, parsing and storing the collected employment data from the information available from the plurality of web sites on the Internet and corresponding URL in conjunction with a processing dictionary that contains a concept based search engine that is configured with documents that contain specific job titles, job descriptions and resume descriptions, into a database;
in the system, automatically updating the employment data stored in the database;
in the system, matching the employment data; and in the system, providing a representative of a non-subscribing entity looking to fill a job position the employment data from the matching, whereby employment needs are fulfilled.

2. The method of claim 1, wherein employment data includes job openings, job postings, job listings, resumes and related employment information.

3. The method of claim 1, wherein the providing includes a non-solicited contacting of the representative.

4. The method of claim 3, wherein the contacting of the representative comprises sending an e-mail.

5. The method of claim 1, further comprising, in the system, making available a placement service.

6. The method of claim 1, further comprising, in the system, revisiting the websites that meet the employment data criteria on a periodic basis to determine whether the content has changed.

7. The method of claim 6, further comprising, in the system, expanding the periodic revisiting time if the content has not changed.