Supplier material market price evaluation method and system based on AI networking search

Abstract

The invention discloses a supplier material market price evaluation method and system based on AI networking search, belongs to the technical field of electronic commerce and data processing, and provides a double-weight confidence model based on time sequence fluctuation and statistical deviation, data are grouped according to platform types and suppliers, a price fluctuation oscillogram is generated for each group of data, and the price fluctuation oscillogram is calculated. Calculating average price and standard deviation in the group; the global average price is introduced, the confidence coefficient of a single group of data is dynamically quantified through a normalization formula, low weight is given to high-fluctuation and high-deviation data, noise interference is effectively suppressed, and the reference of the finally output optimized market price is improved; in addition, structured data with timestamps are generated and the grouping average price and confidence factor of each platform are synchronously output, so that purchasing personnel are assisted to transversely compare the quotation rationality of suppliers; and finally, the output optimized market price is fused with multi-source high-confidence data, so that enterprises are helped to accurately grasp the market price interval, and the bargaining capability and the purchasing efficiency are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of electronic commerce and data processing, and particularly relates to a supplier material market price evaluation method and system based on AI network search. BACKGROUND

[0002] With the rapid development of electronic commerce, the demand for enterprise procurement materials is increasing. The traditional way of obtaining material market price mainly relies on manual inquiry, offline market research, etc., which is low in efficiency and lagging in information. Although some existing online platforms provide certain market price information, the data accuracy and comprehensiveness are insufficient, and there is a lack of intelligent analysis and evaluation. There is no mature solution in the market that can efficiently and accurately integrate AI network search and material market price evaluation and return structured data Currently, enterprises have difficulty in obtaining real-time and accurate supplier material market price information. Manual acquisition method is high in cost and low in efficiency, and it is easy to miss the best procurement opportunity. The data sources of existing online platforms are complex, and the quality is uneven, lacking effective screening and evaluation mechanism, which is difficult to meet the high-precision requirements of enterprises for market price reference. In order to solve the above problems, the present application provides the following technical scheme. SUMMARY

[0003] The purpose of the present application is to provide a supplier material market price evaluation method and system based on AI network search, which solves the problem that the existing technology mainly relies on manual inquiry, offline market research, etc. for enterprise to obtain material market price, which is low in efficiency and lagging in information. Although some existing online platforms provide certain market price information, the data accuracy and comprehensiveness are insufficient, and there is a lack of intelligent analysis and evaluation.

[0004] The purpose of the present application can be achieved by the following technical scheme: The supplier material market price evaluation method based on AI network search comprises the following steps: Step 1, input the key words, and crawl the data resources related to the price of the supplier material on the network; Step 2, data cleaning and preprocessing of data resources; Step 3, grouping the crawled data resources according to the supplier and the data source, each group of data resources obtained from the same supplier and the same type of platform; For a supplier, obtain each group of data resources Zi corresponding to the supplier; Where i takes the value of 1 to m, and m represents the number of data resource sources; Calculate the confidence factor Xi corresponding to each group of data resources; Then according to the formula The optimization market price Ps corresponding to each supplier is calculated, and pppi is the average value of each price data in the set of data resources Zi; Then the structured data is generated; Step 4, continuously track network data dynamics, update the evaluation results in real time, and push the latest data to the target object.

[0005] As a further scheme of the present application, the method for calculating the confidence factor Xi corresponding to each set of data resources is: For a set of data resources Zi, each price data obtained therein is marked in a coordinate system with time as the horizontal coordinate and price unit as the vertical coordinate in chronological order, and a price fluctuation waveform corresponding to the set of data resources Zi is obtained after fitting; The average value pppi of each price data in the set of data resources Zi is calculated; The standard deviation Si of the corresponding price data in each set of data resources Zi is calculated; The average value PP of the corresponding m pppis is calculated; After normalizing each standard deviation Si and |PP-ppi|, the weight values of each set of data resources corresponding to the standard deviation Si and |PP-ppi| are assigned, and the sum of the weight values is obtained to obtain the confidence parameter of each set of data resources; After normalizing the m confidence parameters Ei obtained by calculation, the confidence factor Xi corresponding to each set of data resources is obtained; As a further scheme of the present application, for a set of data resources, the confidence parameter Ei=1 / (a1*Si+a2*|PP-ppi|), wherein a1 and a2 are both preset coefficients, and a1+a2=1.

[0006] As a further scheme of the present application, the method for preprocessing the data resources in step 2 is: Identify the key information in the text; After data deduplication, unit unification is performed, and missing values are filled or discarded.

[0007] As a further scheme of the present application, the method for data cleaning of the data resources in step 2 is: Set a reasonable price range, and remove all price data exceeding the range; or Use Z-score or IQR statistical method to identify and remove abnormal data.

[0008] As a further scheme of the present application, in step 1, NLP technology is also used to analyze the material keywords, automatically extract synonyms, near synonyms, related terms, common aliases, and model variants, and when crawling data resources, these expanded words are also used as keywords.

[0009] As a further scheme of the application, the data is updated and pushed every preset time length t1, and the data resources crawled each time are from data resources within a past preset time length t2 starting from the current time.

[0010] The application further discloses a supplier material market price evaluation system based on AI network search. The AI network search module crawls and filters massive network resources through keywords, and locks web pages, forums and e-commerce platforms containing supplier material price information. The data cleaning and preprocessing module unifies the format, removes duplicates and invalid information, and extracts key fields such as material variety, model, specification, unit, price and price update date.

[0011] The market price evaluation algorithm module analyzes the cleaned data, calculates the optimized market price Ps after evaluation, and generates structured data. The data updating and pushing module monitors network data changes in real time, updates evaluation results regularly, and pushes the latest structured data to users or systems in need.

[0012] The application has the following advantages: The application proposes a double-weight confidence model based on time series fluctuation and statistical deviation, groups data according to platform types (B2B / government / second-hand market, etc.) and suppliers, generates price fluctuation waveform graphs for each group of data, calculates the average price ppi and the standard deviation Si in each group, introduces the global average price PP, dynamically quantifies the confidence of single group data through a normalization formula, gives low weight to data with high fluctuation and high deviation, effectively suppresses noise interference, and improves the referenceability of the optimized market price output. In addition, the application generates structured data with timestamps and synchronously outputs the average price ppi of each platform group and the confidence factor Xi, which helps procurement personnel compare the rationality of supplier quotes horizontally. The optimized market price Ps output finally fuses multiple high-confidence data, helps enterprises accurately grasp the market price interval, and improves the bargaining power and procurement efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0013] The application will be further described below with reference to the accompanying drawings.

[0014] Figure 1 FIG. 1 is a flowchart of a supplier material market price evaluation method based on AI network search of the application; Figure 2 FIG. 2 is a schematic diagram of the framework structure of a supplier material market price evaluation system based on AI network search of the application. DETAILED DESCRIPTION

[0015] With reference to the drawings and embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0016] Embodiment 1 A supplier material market price evaluation method based on AI network search, as shown in Figure 1 The method comprises the following steps: Step 1, first, start the AI network search module, input the keywords, and crawl the data resources such as web pages, posts, and commodity pages related to the supplier material price on the network; The keywords include material name, specification and model, brand (supplier), industry standard, etc. Further, the NLP technology can be used to analyze the material keywords, automatically extract synonyms, near synonyms, related terms, common aliases, and model variants (such as 304 stainless steel can be expanded to SUS304, A2 stainless steel). When crawling data resources, these expanded words are also used as keywords; Step 2, data cleaning and preprocessing module is used to clean and preprocess the data resources; The method for cleaning and preprocessing the data resources is as follows: Preprocessing: Identify the material name, specification and model, price value, currency unit, quantity unit, date, company name, and other key information in the text; After data deduplication, unify the units and fill or discard the missing values; The method for deduplication is to identify and merge duplicate quotes from the same supplier or the same commodity page; Unit unification refers to the unification of the units related to the same index, such as: All prices are converted into the target currency (such as RMB CNY, USD) and the target unit (such as / kg, / piece, / ton), and if the exchange rate conversion is involved, the real-time or historical exchange rate and unit conversion are used, and different specifications (such as 50*50*2mm, 50mm x 50mm x 2mm, 5cm square tube wall thickness 2mm) are mapped to the unified standardized format.

[0017] Filling or discarding missing values means: For records with missing key fields (such as price, unit), fill or discard them according to the corresponding strategy (such as using the average price of the same specification material).

[0018] Cleaning: The abnormal price data is removed, and the specific removal method is as follows: set a reasonable price range, and delete all price data exceeding the range; Or use statistical methods such as Z-score or IQR to identify and remove several abnormal data; Step 3, classify the data resources crawled by the AI networking search module, specifically according to the source of the data resources; For example: divide the data resources into B2B platform data, government / association platform data, news media platform data, industry vertical website data, auction / second-hand platform data, etc.; Then further group the data resources according to different brands, so that each group of data resources obtained is from the same supplier and the same type of platform; For a supplier, obtain each group of data resources corresponding to the supplier, and mark each group of data resources corresponding to the supplier as Zi; Where i takes the value of 1 to m, and m represents the number of types of data resource sources; For a group of data resources Zi, each price data obtained in the group is marked in the coordinate system with time as the horizontal coordinate and price unit as the vertical coordinate in chronological order, and a price fluctuation waveform corresponding to the group of data resources Zi is obtained after fitting; Further calculate the average value ppi of each price data in the group of data resources Zi; Calculate the standard deviation Si of the corresponding price data in each group of data resources Zi; Calculate the average value PP of the corresponding m ppi; After normalizing each standard deviation Si and |PP-ppi|, the weight value of each group of data resources corresponding to the standard deviation Si and |PP-ppi| is assigned, and the sum of the weight values is obtained. The confidence parameter corresponding to each group of data resources is obtained; Specifically, for a group of data resources, its confidence parameter Ei=1 / (a1*Si+a2*|PP-ppi|), where a1 and a2 are both preset coefficients, and a1+a2=1; After normalizing the m confidence parameters Ei obtained by calculation, the confidence factor Xi corresponding to each group of data resources is obtained; Then according to the formula The optimized market price Ps corresponding to each supplier is calculated.

[0019] Then generate structured data, such as: [{ "material": <Raw material type including model, specification and unit>, "price": <Raw material type price>, "date": <Latest price date>}]; In addition to optimizing the market price Ps and its corresponding material name, specification, and model data, the output structured data can also include the corresponding average ppi on different types of platforms and the material name, specification, and model data corresponding to each ppi, making it easier for staff to make reasonable judgments about supplier materials. Step 4: Through the data update and push module, continuously track network data dynamics, ensure that the evaluation results are updated in real time, and push the latest data to the target objects in a timely manner.

[0020] Specifically, the data can be updated and pushed every preset time period t1, and the data resources crawled each time come from the data resources within a certain preset time period t2 in the past starting from the current moment, so as to avoid the time span being too large, resulting in data distortion, loss of reference significance, and affecting the accuracy of the final output structural data.

[0021] Example 2 This embodiment discloses a supplier material market price evaluation system based on AI network search, such as Figure 2 As shown, the system includes: AI online search module: crawls and filters massive network resources through intelligent algorithms, locking in web pages, forums, e-commerce platforms, etc. that contain supplier material price information to ensure that the data source is extensive and authoritative.

[0022] Data cleaning and preprocessing module: unify the format of the collected raw data, remove duplicates, remove invalid information, and extract key fields such as material varieties, specifications, units, prices, and price update dates.

[0023] Market price evaluation algorithm module: Analyzes the cleaned data based on a specific algorithm, comprehensively considers factors such as quotations from different suppliers, market supply and demand, and regional differences, calculates the optimized market price Ps after evaluation, and generates structured data, such as: [{ "material": <Raw material type including model, specification and unit>, "price": <Raw material type price>, "date": <Latest price date>}] Data update and push module: monitors network data changes in real time, regularly updates assessment results, and proactively pushes the latest structured data to users or systems in need.

[0024] The above merely illustrates and describes the present application, and those skilled in the art can make various modifications or supplements to the described specific embodiments or adopt similar ways to replace, as long as the modifications or supplements do not deviate from the present application or exceed the scope defined by the present application, and should belong to the protection scope of the present application.

Claims

1. The supplier material market price evaluation method based on AI network search is characterized by: The steps include: Step 1: Enter keywords and crawl data resources related to supplier material prices on the Internet; Step 2: Clean and preprocess the data resources; Step 3: Group the crawled data resources according to suppliers and data sources. Each group of data resources obtained comes from the same supplier and the same type of platform. For a supplier, obtain the corresponding data resources Zi; The value of i ranges from 1 to m, and m represents the number of source types of data resources; Calculate the confidence factor Xi corresponding to each group of data resources; Then according to the formula Calculate the optimized market price Ps corresponding to each supplier; ppi is the average value of each price data in a set of data resources Zi; Then generate structured data; Step 4: Continuously track network data dynamics, update assessment results in real time, and push the latest data to the target object.

2. The supplier material market price evaluation method based on AI network search according to claim 1 is characterized in that: The method for calculating the confidence factor Xi corresponding to each group of data resources is: For a set of data resources Zi, the price data obtained therein are marked in chronological order in a coordinate system where the horizontal axis is time and the vertical axis is price unit. After fitting, a price fluctuation waveform corresponding to a set of data resources Zi is obtained; Calculate the average ppi of each price data in this set of data resources Zi; Calculate the standard deviation Si of the price data corresponding to each group of data resources Zi; Calculate the average value PP of the corresponding m ppi; After normalizing each standard deviation Si and |PP-ppi|, assign weight values ​​to the standard deviation Si and |PP-ppi| corresponding to each group of data resources and sum them up to obtain the confidence parameters corresponding to each group of data resources; After normalizing the m confidence parameters Ei obtained by calculation, the confidence factor Xi corresponding to each group of data resources is obtained.

3. The supplier material market price evaluation method based on AI network search according to claim 2 is characterized in that: For a set of data resources, the confidence parameter Ei=1 / (a1*Si+a2*|PP-ppi|), where a1 and a2 are both preset coefficients, and a1+a2=1.

4. The supplier material market price evaluation method based on AI network search according to claim 1 is characterized in that: The method for preprocessing data resources in step 2 is: Identify key information in a text; After deduplication, the data is normalized and missing values ​​are filled or discarded.

5. The supplier material market price evaluation method based on AI network search according to claim 1 is characterized in that: The method for data cleaning of data resources in step 2 is: Set a reasonable price range and remove all price data that exceeds the range; or Use Z-score or IQR statistical methods to identify abnormal data and then eliminate them.

6. The supplier material market price evaluation method based on AI network search according to claim 1 is characterized in that: In step 1, NLP technology is also used to analyze material keywords, extract synonyms, antonyms, related terms, common aliases, and model variants, and these expanded words are also used as keywords when crawling data resources.

7. The supplier material market price evaluation method based on AI network search according to claim 1 is characterized in that: The data is updated and pushed every preset time t1, and the data resources crawled each time come from the data resources within the preset time t2 in the past starting from the current moment.

8. A supplier material market price evaluation system based on AI network search, used to implement the evaluation method according to any one of claims 1 to 7, the system comprising: The AI ​​network search module crawls and filters massive network resources using keywords, targeting web pages, forums, and e-commerce platforms that contain supplier material price information; The data cleaning and preprocessing module unifies the format of the collected raw data, removes duplicates, and removes invalid information, extracting key fields such as material types, specifications, units, prices, and price update dates; The market price evaluation algorithm module analyzes the cleaned data, calculates the optimized market price Ps after evaluation, and generates structured data; The data update and push module monitors network data changes in real time, regularly updates evaluation results, and pushes the latest structured data to users or systems in need.

Images