Search

KR-102676472-B9 - METHOD FOR PROVIDING ELECTRONIC BIDDING INFORMATION ANALYSIS SERVICE USING BIDSCORE FOR PROCUREMENT AUCTION

KR102676472B9KR 102676472 B9KR102676472 B9KR 102676472B9KR-102676472-B9

Abstract

A method for providing an electronic bidding information analysis service using a bid score is provided, comprising the steps of: building a winning bid information database by collecting data including at least one participating company name and multiple preliminary prices from at least one winning bid result; receiving an announcement number to participate in the bid from a user terminal; collecting at least one participating company name that participated in the announcement number; extracting multiple preliminary prices that are already mapped and stored to at least one participating company name to load 1,365 multiple preliminary prices; and setting the number of multiple preliminary prices obtained from the user terminal among the 1,365 multiple preliminary prices as a bid score and transmitting it to the user terminal.

Inventors

  • 조충환

Assignees

  • 주식회사 비드스코어

Dates

Publication Date
20260507
Application Date
20211022

Claims (10)

  1. In a method for providing analysis services executed on an analysis service providing server, A step of constructing a winning bid information database by collecting data including at least one participating company name and multiple preliminary prices from at least one winning bid result; A step of receiving an announcement number to be entered from a user terminal to participate in the bid; A step of collecting the name of at least one participating company that participated in the above announcement number; A step of extracting multiple preliminary prices that are already mapped and stored to at least one participating company name and loading 1,365 multiple preliminary prices; A step of setting the number of multiple reserve prices obtained from the user terminal among the above 1,365 multiple reserve prices as a bid score and transmitting it to the user terminal; When a winning bid announcement is uploaded, a step of adding a bid score column to the winning bid announcement detail page to display the bid score; A step of generating feedback data by comparing the above-mentioned output bid score and winning bid ranking; and A step of predicting the Reservation Price calculated from the above multiple reserve prices using at least one artificial intelligence algorithm that has been trained and tested on big data built in the above winning bid information database, Includes, The above-mentioned at least one artificial intelligence algorithm is an MLP (Multi Layer Perceptron) algorithm and an ANFIS (Adaptive Neural Fuzzy Inference System) algorithm, which is an adaptive neural fuzzy inference system that combines the IF-THEN concept of FIS (Fuzzy Inference System) and the learning ability of ANN. The above ANFIS algorithm is, It is defined as a fuzzy inference system in which the parameter values of the membership function are adjusted based on a given input and output data set, using the backpropagation algorithm or a combination of backpropagation and least squares, which is a hybrid-learning algorithm, It consists of five layers, including Layer 1, which calculates the membership values of the antecedent variables as a fuzzification layer; Layer 2, which calculates the influence of the Rule; Layer 3, which normalizes the strength of the influence; Layer 4, which adjusts the consequent parameters (pv, qv, rl); and Layer 5, which calculates the total output value. The step of predicting with at least one artificial intelligence algorithm is, For the modeling of the above artificial intelligence algorithm, construction winning bid prices in each field among past electronic bidding data are collected, and winning bid information data ordered through the Public Procurement Service's KONEPS over the past 5 to 10 years is used as a sample; the above modeling is performed through the process of data collection, data preprocessing, data modeling, comparison of bid amount prediction and regression indicator-based models, and model evaluation regarding data from public and private ordering agencies. After collecting and preprocessing bid data from specialized bidding sites and the Public Procurement Service, four characteristics including the base price, estimated price, range of multiple preliminary prices, and lower limit of the winning bid rate are set as input variables, and the winning bid amount is set as the output variable, After training with Python and TensorFlow using a 7:3 ratio of training to test data, designate the winning bid data from the past 3 months as the actual data, Subsequently, since the winning bid amount data is large, the above winning bid amount data is scaled using Min-Max (Min-Max Normalization) and Std (Standardization), and then prediction experiments are conducted using deep learning techniques, specifically the MLP algorithm and the ANFIS algorithm model. Set the epoch for each model, change the learning rate, and conduct repeated experiments N or more times, and In the ANFIS algorithm, regarding the number of fuzzy rules, the maximum number of control rules is set to 16 as there are 4 input variables (control variables) and 2 fuzzy language variables, and When performing the above model evaluation, the algorithm models are compared using regression indicators of Mean Squared Error (MSE) expressed by Equation 1 below, Root Mean Square Error (RMSE) expressed by Equation 2 below, Mean Absolute Error (MAE) expressed by Equation 3 below, and Mean Absolute Ratio Error (MAPE) expressed by Equation 4 below, based on the predicted and actual values. [Mathematical Formula 1] [Mathematical Formula 2] [Mathematical Formula 3] [Mathematical Formula 4] In the above mathematical formulas 1 to 4, n is the number of data points, and yi is the actual observed value, is the predicted value, ε is the error, and The best learning rate is found by comparing the results of the above MLP algorithm and the above ANFIS algorithm, a high-accuracy algorithm is applied where the difference between the actual value and the predicted value is not large, and continuous retraining is performed using the above feedback data, A method for providing an electronic bidding information analysis service using a bid score, which is possible to improve the performance of winning bid price prediction and assist each participating company in making quick decisions through prediction using the above algorithm.
  2. In Article 1, The above user terminals are multiple user terminals of different participating companies, and After the step of receiving the announcement number to be entered from the user terminal, A step of securing multiple preliminary prices entered by the plurality of user terminals for the bid of the above announcement number, and counting the number of participating companies of the plurality of user terminals; A step of calculating the average bid score by dividing 1365 by the number of the aforementioned counted participating companies; A step of setting the number of multiple reserve prices held by the first-ranked participating company that matches the largest number of multiple reserve prices among the above 1,365 multiple reserve prices as the first-ranked bid score; A step of visualizing the above average bid score and the above first-rank bid score; A method for providing an electronic bidding information analysis service using a bid score, characterized by further including
  3. delete
  4. In Article 2, After the step of setting the number of multiple reserve prices obtained from the user terminal among the above 1365 multiple reserve prices as the bid score and transmitting it to the user terminal, A step of guiding the user terminal to increase the probability of winning the bid by maintaining a bid score of 1.5 to 2 times the average bid score; A method for providing an electronic bidding information analysis service using a bid score, characterized by further including
  5. In Article 2, In the step of securing multiple preliminary prices entered by the multiple user terminals for the bid of the above announcement number and counting the number of participating companies of the multiple user terminals, A method for providing an electronic bidding information analysis service using a bid score, characterized by using a business registration number as an identification code when distinguishing the above-mentioned participating companies.
  6. delete
  7. delete
  8. In Article 1, After the step of setting the number of multiple reserve prices obtained from the user terminal among the above 1365 multiple reserve prices as the bid score and transmitting it to the user terminal, A step of predicting the probability distribution of the estimated price calculated from the above multiple preliminary prices by analyzing descriptive statistics between at least one pre-set independent variable, performing correlation analysis, and then predicting the probability distribution of the estimated price using an analysis technique of inverse probability-based modeling; A method for providing an electronic bidding information analysis service using a bid score, characterized by further including
  9. In Article 1, The above user terminals are multiple user terminals of different participating companies, and After the step of setting the number of multiple reserve prices obtained from the user terminal among the above 1365 multiple reserve prices as the bid score and transmitting it to the user terminal, A step of visualizing the distribution of multiple preliminary prices collected from the above multiple user terminals and displaying different participating companies, which are multiple users, in the fraudulent multiple preliminary prices by pseudonymizing them: A method for providing an electronic bidding information analysis service using a bid score, characterized by further including
  10. In Article 1, A method for providing an electronic bidding information analysis service using a bid score, characterized in that the number of multiple reserve prices obtained from the user terminal is the number of multiple reserve prices that matches the 1,365 multiple reserve prices.

Description

Method for Providing Electronic Bidding Information Analysis Service Using BidScore The present invention relates to a system for providing an electronic bidding information analysis service using a bid score, and provides a method for providing to a user a number of multiple reserve prices secured by the user among multiple reserve prices by setting it as a bid score. In public construction projects funded by state resources, the estimated price serves as the basis for contract execution and significantly influences the assurance of project quality and the revitalization of the contractor's business conditions. An estimated price refers to a value prepared and kept on file in advance prior to bidding or contract signing to serve as a standard for determining the successful bidder and the contract amount, specifically the price established pursuant to Article 8 of the Enforcement Decree of the State Contract Act. There are two methods for determining the estimated price: the single preliminary price method and the multiple preliminary price method. The multiple preliminary price method involves generating multiple preliminary prices based on a base amount and determining the estimated price by calculating the arithmetic mean of a few selected preliminary prices. Currently, Korean government agencies, the Public Procurement Service, and public institutions apply the multiple preliminary price method rather than the single preliminary price method. The multiple preliminary price system is implemented to prevent prior knowledge of the estimated price, as knowing the price in advance offers an absolute advantage in winning the bid. At this time, research and development were conducted on methods for automatically bidding in the public procurement market by analyzing electronic bidding methods and big data. In this regard, prior art Korean Patent Publication No. 2003-0030140 (published April 18, 2003) and Korean Patent Publication No. 2017-0069907 (published June 21, 2017) disclose configurations that simultaneously process electronic bidding and document bidding using the Internet in e-commerce between the government, public institutions, and companies, enable the automatic transfer of bidding-related fees and bid security deposits through authentication by an accredited certification authority, and allow the bidding process conducted by the Public Procurement Service—including bidding information, specification drafting, joint venture agreements, bid documents, and announcements—to be executed through authentication and security. Additionally, configurations are disclosed that collect data from a bidding server, calculate the bidding time and predicted bid price based on big data when option information is selected on a user terminal, and provide an estimated bid price. However, the former holds no significance beyond converting offline bidding to electronic bidding, and the latter, composed of predictions regarding advertising exposure rankings or bid prices, is far removed from public procurement bidding. Prior to the 2000s, contract managers created 10 preliminary prices and determined the estimated price by calculating the arithmetic mean of three randomly selected from them on the day of bidding. However, since 2000, with the advancement of information technology, the Public Procurement Service and various ordering agencies have established their own electronic bidding systems, allowing contract managers to automatically calculate the estimated price through these systems. Most ordering agencies adopt a method where 15 preliminary prices are randomly selected within a certain range centered on a base amount, and the estimated construction price is determined by calculating the arithmetic mean of four values randomly drawn from among them on the day of bid opening. Accordingly, research and development of a platform capable of counting the number of matching multiple preliminary prices among participating companies and providing this as a bid score is required. FIG. 1 is a diagram illustrating a system for providing an electronic bidding information analysis service using a bid score according to an embodiment of the present invention. Figure 2 is a block diagram illustrating an analysis service provider server included in the system of Figure 1. FIGS. 3 and 4 are drawings for explaining an embodiment in which an electronic bidding information analysis service using a bid score according to an embodiment of the present invention is implemented. FIG. 5 is an operation flowchart illustrating a method for providing an electronic bidding information analysis service using a bid score according to an embodiment of the present invention. Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments d