A Survey Of E-Tendering And Use Of Nlp In Automated Tender Evaluation

2531 (6 pages)
Download for Free
Watch out! This text is available online and is used for guidance and inspiration
Download PDF

I. Introduction

Natural Language Processing is a branch of Artificial Intelligence which enables computers to analyse and understand the human language. Natural Language Processing (NLP) was formulated to build software that generates and understand natural languages so that a user can have natural conversations with his computer. NLP combines AI with computational linguistics and computer science to process human or natural languages and speech. [10]In general terms, the objective of NLP is to exploit the natural languages as successfully as man does. Natural language, regardless of whether talked, composed, or typed, is the most normal method for interaction among people, and the frequently chosen method of articulation for a large portion of the reports they create. [15]

As PCs assume a bigger part in the readiness, securing, transmission, checking, capacity, examination, and change of data, investing them with the capacity to comprehend and produce data communicated in natural language turns out to be increasingly essential. Procurement Management Process, is a method by which items are purchased from external suppliers. [13] The procurement management process involves managing the ordering, receipt, review and approval of items from suppliers. Under procurement management we find tendering, which is defined as making a formal written offer to carry out work, supply goods, or buy land, shares, or another asset for a stated fixed price. [14] The tender process in Zimbabwe is such that tenders are being posted in newspapers and suppliers have to submit applications in the form of enveloped letters many at times directly to the organization offering the tender. This is resource consuming and fails to capture all of the potential suppliers. To save time and safeguard the process from human error the researcher is focusing on automating the tender process using NLP for analysis and evaluation of vendors.

There is therefore need to criticize existing NLP methods and approaches in order to identify the most appropriate technology for the project. The qualified technology should be relevant, accurate and specific.

II. Related work

In this section, the researcher endeavors to decide the present state-of-play in connection to e-Tendering opportunities, by inspecting a few International e-Tender frameworks (usefulness, capacity, and so forth) recognized amid the research. MERX eTendering Defined as Canada’s official public-sector electronic tendering service, allows construction industry businesses to have easy and affordable access to billions of dollars in contracting opportunities with the federal government and participating provincial and municipal governments. With billions of dollars in public sector business opportunities tendered annually, suppliers to the government for example, can use MERX eTendering to connect to buyers in the federal, provincial and MASH (municipal, academic, school and hospital) sectors to get the information they need to bid on public sector contracts. In addition to having the ability to search for open tender opportunities from the more than 1,500 opportunities available at any time, the MERX eTendering service reportedly has a comprehensive document delivery service that maintains high service levels for delivering tender documents.

Bid ExpressAlthough not an electronic ‘tendering’ system per se, the Bid Express system was included in this investigation due to its unique electronic ‘bidding’ process. Bid Expressis a web-based bidding information service developed exclusively for the USA highway construction industry to increase the efficiency and accuracy of the existing bidding process – that is, by saving time traditionally required for preparing bids on paper, as well as needless travel time and expense to attend lettings or submit bids in person. Bid Express has been in operation for over five years, allowing registered members to scan all current lettings, historical data from previous lettings, and vendor information using a keyword search facility. In addition, members have instant access to lettings from across the country, not just those of their home state. When bidding electronically using Bid Express, there is no delay in learning the outcome. This is due to the state transportation agency having the option of posting the results onto the Bid Express web site, as they are being read – that is, ‘real-time’ bid results available within seconds of the bid opening.

Contractors also have the flexibility to withdraw or replace their bids at any time prior to opening (Bid Express 2003). TenderTrustIn April 2001, the UK Office of Government Commerce (OGC) announced the start of a six- month trial of the online tendering system TenderTrust (Figure 4-21). Through this cross- government initiative, it was believed that the new system, if it passes its pilot stage, should produce a savings of about ǧ13 million for the UK taxpayer, and reduce annual tendering costs for suppliers by ǧ37 million (E-Government Bulletin 2001). In 2001, the TenderTrust system was identified as the world’s first smartcard internet-based eTendering system. In association with the Royal Bank of Scotland, the system used ‘banking strength’ digital certificates (Public Key Infrastructure technology and encryption techniques) to meet the highest security and authentication standards throughout the entire tendering process (Byatt I. 2001).

As such, the Tender

Trust system was specifically designed to meet the demanding requirement of both public and private sector tendering, and to assist in removing many of the inefficiencies in traditional paper-based tendering (Gershon P. 2001). [9]DELTA e-Tendering SuiteThis was introduced into the United Kingdom council’s procurement process to reduce administration costs and burdens for both buyers and suppliers. It is used to effortlessly create, manage and transmit contract announcements electronically, enterprise-wide, regardless of location. I has an online supplier information database listing service that provide public sector Buyers with instant access to up-to-date data on pre-registered Suppliers. Provides an internet-based secure tender box facility for Buyers and Suppliers to exchange documents electronically – that is, an electronic e-tendering service that provides the procurement community (Buyers & Suppliers) a secure, cost-effective and easy-to-use e-tendering solution for the transmission, retrieval, storage, receipt and administration of Invitation to Tender (ITT) documentation. [11]ELTONSuffolk City Council’s Procurement and Commissioning department, as part of their commitment to making communication with the County Council easier and faster, have developed ELTON (Electronic Tendering Online), thereby allowing the council to issue tenders electronically and companies to submit bids and proposals via e-mail. European Union (EU) regulations, Acts of Parliament and the council’s own internal standing orders govern the entire tender process, thereby prescribing how the tender is advertised, specified, submitted and evaluated. Most contracts are advertised via the above website, as well as in the local press, trade publications and the European Union (EU) Journal where appropriate.

The contract period varies (anything from a one-off tender to over 5 years). Applications for tender documents can be sent by post, email or fax. For contracts over a certain value, ELTON will issue a pre-qualification (PQC) questionnaire prior to issuing the invitation to tender in order to save potential bidders, who do not meet the selection criteria, time and effort in completing tender documents. Tender documents are to be completed and returned by the date specified in the document, and any tender received late will not be considered (regardless of when it was sent). Returned tenders (usually via email attachments) are evaluated against pre-determined criteria (usually based on an acceptable level of both quality and price), after which tenderers will receive a letter informing them whether their bid was successful or not. [12]NLP approaches

A. Automatic Summarization

Automatic Summarization is a technique that reduces a large text into a meaningful short paragraph. This approach retrieves the summary of large text documents. Internet search engines use this technique to identify the content of a website so that can be indexed appropriately. In general, there are two types of approaches used in Automatic Summarization. One is Extraction which extracts few important parts from the text such as key word sentences or paragraphs [1]. Another type of approach is Abstraction which paraphrases the important points of the text. Technically, Abstraction is a more complicated system than Extraction to develop. Mainly, each Auto Summarization system’s functionality can be divided into three general steps[1].

We will write a unique paper on this topic for you!
Place Order

*No hidden charges

Analysis: In the analysis stage, the text is analysed and an internal representation is generated. This internal representation is produced in the way logical relationships between sentences can be established.

Transformation: In Transformation stage, the internal representation is manipulated to produce an ordered text representation. In general, the ordered representation could have sentences ranked by a scoring function. The scoring method is usually based on the facts found in analysis stage. A simple example is where sentences with frequent keywords get a specific score.

Realization: In Realization stage, the summary is generated based on the scoring of Transformation stage. In some systems it could be simply producing specific scored sentences from the text. There are many auto summarizing tools available. smmry. com is an online tool that produces summaries either by providing URL of a web page or uploading a file. smmry. com also provides API for web developers who could use auto summarization for their developments. [24]

B. Named Entity Recognition

Named Entity Recognition is the subject area that identifies entities or physical objects such as names of persons, organizations, and places in the text. For an example, Named Entity recognition of sports news would consist of names of the players, places of teams and grounds, etc. Traditionally, grammar based approaches are used to identify name and entities in the text. In the present systems, statistical models are incorporated in order to classify names and entities more precisely. In statistical approaches, initially, a set of training data will be used against the model. Based on this training data, statistics will be prepared. These statistics will be used against real documents [3]. The advantage of statistical approach is it can be continuously improved as it is used on various texts. Historical data patterns will be used in identifying new name entities. The Natural Language Processing Group at Stanford University offers variety of solutions to NLP problems. Stanford NER is a Named Entity Recognizer which is implemented in Java platform and provides libraries with a way to identify names and entities [4].

C. Part-of-speech tagging

In this NLP approach, sentences will be tagged according to the grammatical order such as nouns, verbs and adjectives. Because of the real nature of a language, some words can have multiple tags such as both noun and verb. Part-of-Speech (POS) tagger programs use combinations of several techniques such as lexicons, rules, and dictionaries [5]. Dictionaries contain categories of words. Usually, tagging programs accurately tag the word or make a best guess. When the words are ambiguous in a sentence, POS taggers use probability approaches to tag correctly.

D. Word-sense Disambiguation

Word-sense Disambiguation is a NLP subject that identifies the correct sense of the word in a sentence where that word could have multiple meanings. This area is an important part in information retrieval because the contextual meaning of the text depends on the correct interpretation of that text. Human languages generally have many ambiguities. A human can understand the meaning of a word based on the context it is spoken and the background knowledge of the subject. However, a machine would have difficulties in identifying such correct meaning. Word-sense disambiguation methods help a machine to minimize the ambiguities of words in the text. In general, a word-sense disambiguation approach consists of a lexical repository that contains different senses for words [6]. WordNet is a free lexical database in English that contains a large collection of words and senses [7]. The design of WordNet was inspired by the theories of human linguistic memory [8]. WordNet encloses a large volume of nouns, verbs, adjectives and adverbs in English language. In WordNet, words are grouped and interlinked by their meanings. This method allows the identification of any closer or similar meaning to a given word. Many NLP applications use WordNet as a source for processing.

E. Sentiment Analysis

Sentiment Analysis is an NLP process which identifies the attitude or contextual polarity of the writer with respect to the text. A text collection could show one or many sentiments. Sentiments can be positive, negative or neutral. Sentiment Analysis is heavily used in processing online reviews for products, movies and books. Many websites run specific algorithms to rate the reviews based on the degree of positive or negative tone. In general, an opinion consists of two components; a target and a sentiment towards the target [9]. Usually, sentiment analysis is conducted after the text is parsed by Part-of-Speech tagging. Specific rules will be applied along with an internal dictionary to POS tags to identify the sentiment. A simple example of a sentiment rule is if two adjoining positive words are adverb and adjective then it is classified as positive sentiment. There are more sophisticated algorithms that exist in order to extract the sentiment. The main use of NLP in the e-tendering project will be to fetch text submitted by the suppliers for evaluation. It will also be used to fetch documents given that the suppliers will be submitting documents. This will make use of the parsing techniques that have been established in NLP. These are basically divided into two, top-down and bottom-up parsing. Giving parsing a definition, it is the process of automatically building syntactic analyses of a sentence in terms of a given grammar and lexicon. The resulting syntactic analyses may be used as input to a process of interpretation and evaluation.

III. Discussion

Although most of the e-Tender websites investigated at the time, maintain their tendering processes and capabilities are ‘electronic’, research shows these ‘eTendering’ systems vary from being reasonably advanced to more ‘basic’ electronic tender notification and archiving services for various industry sectors. Research also indicates an e-Tender system should have a number of basic features and capabilities, including:

  • All tender documentation to be distributed via a secure web-based tender system – thereby avoiding the need for collating paperwork and couriers.
  • The client/purchaser should be able to upload a notice and/or invitation to tender onto the system.
  • Notification is sent out electronically (usually via email) for suppliers to download the information and return their responses electronically (online).
  • During the tender period, updates and queries are exchanged through the same e-Tender system.
  • The client/purchaser should only be able to access the tenders after the deadline has passed.
  • All tender related information is held in a central database, which should be easily searchable and fully audited, with all activities recorded.
  • It is essential that tender documents are not read or submitted by unauthorized parties.
  • Users of the e-Tender system are to be properly identified and registered via controlled access. In simple terms, security has to be as good as if not better than a manual tender process. Data is to be encrypted and users authenticated by means such as digital signatures, electronic certificates or smartcards.
  • All parties must be assured that no ‘undetected’ alterations can be made to any tender.
  • The tenderer should be able to amend the bid right up to the deadline – whilst the client/purchaser cannot obtain access until the submission deadline has passed.
  • The e-Tender system may also include features such as a database of service providers with spreadsheet-based pricing schedules, which can make it easier for a potential tenderer to electronically prepare and analyze a tender.

The literature indicate the feasibility of the automated evaluation of bids sent by vendors. Moreover, the high correlation between manual and automated evaluations suggests that the less labor-intensive automated evaluations can be used as a proxy for human evaluations when developing e tendering systems. [25] This is of significant practical value.

You can receive your plagiarism free paper paper on any topic in 3 hours!

*minimum deadline

Cite this Essay

To export a reference to this article please select a referencing style below

Copy to Clipboard
Writingbros. (2020, July, 15) A Survey Of E-Tendering And Use Of Nlp In Automated Tender Evaluation. Retrived September 19, 2020, from https://writingbros.com/essay-examples/a-survey-of-e-tendering-and-use-of-nlp-in-automated-tender-evaluation/
"A Survey Of E-Tendering And Use Of Nlp In Automated Tender Evaluation." Writingbros, 15 Jul. 2020, https://writingbros.com/essay-examples/a-survey-of-e-tendering-and-use-of-nlp-in-automated-tender-evaluation/. Accessed 19 September 2020.
Writingbros. 2020. A Survey Of E-Tendering And Use Of Nlp In Automated Tender Evaluation., viewed 19 September 2020, <https://writingbros.com/essay-examples/a-survey-of-e-tendering-and-use-of-nlp-in-automated-tender-evaluation/>
Writingbros. A Survey Of E-Tendering And Use Of Nlp In Automated Tender Evaluation. [Internet]. July 2020. [Accessed September 19, 2020]. Available from: https://writingbros.com/essay-examples/a-survey-of-e-tendering-and-use-of-nlp-in-automated-tender-evaluation/
Copy to Clipboard

Need writing help?

You can always rely on us no matter what type of paper you need

Order My Paper

*No hidden charges