Cross-platform information comparison and consumption decision support system and method based on mobile phone terminal

Abstract

The invention relates to a cross-platform information comparison and consumption decision support system and a cross-platform information comparison and consumption decision support method based on a mobile phone terminal, which realize multi-dimensional information integration and intelligent decision support by constructing a'consumer-physical store-platform 'triangular interactive ecology. A dual-track data acquisition technology with API priority is adopted, and mainstream platform data are synchronously captured at a millisecond level; a dynamic data standardization protocol is integrated with BERT semantic matching and an XGBoost false promotion recognition algorithm to solve the naming heterogeneous problem. The multi-dimensional comparison model supports 20 + index analysis of price, service quality, timeliness and the like. Transparent decision support is provided for consumers through an interface-OCR-crawler three-level data acquisition mechanism, the decision time is shortened by 55% through actual measurement of thousands of users, and the repurchase rate of merchants is increased to 35%. The system strictly follows the GDPR and individual insurance law compliance requirements, all data acquisition needs to be authorized item by item by a user, and the system has remarkable social and economic benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of consumer decision-making, and specifically provides a cross-platform information comparison and consumer decision-making support system and method based on a mobile phone. Background Art

[0002] The existing consumer decision-making support systems have the following technical defects:

[0003] 1. The phenomenon of information silos is serious. Users need to frequently switch between different APPs to compare prices, and the time-consuming for a single decision-making is up to 3-5 minutes;

[0004] 2. The data heterogeneity is prominent. The naming rules of goods and price calculation methods on each platform are significantly different, resulting in the information integration accuracy rate being less than 65%;

[0005] 3. Lack of a dynamic monitoring mechanism. It is difficult for users to capture real-time price fluctuations (the average price change response delay exceeds 4 hours);

[0006] 4. Physical stores are overly dependent on platform traffic. The commission cost accounts for up to 20-35%, and they cannot reach consumers independently;

[0007] 5. The consumer decision-making dimension is single. Traditional systems only provide price comparison and lack in-depth analysis indicators such as service quality and platform credit.

[0008] The above problems lead to low consumer decision-making efficiency, high business operation costs, and insufficient market information transparency. There is an urgent need to build an innovative solution for cross-platform data integration and intelligent decision-making support Summary of the Invention

[0009] In order to solve the problems of the existing technology, the present invention provides a cross-platform information comparison and consumer decision-making support system and method based on a mobile phone.

[0010] In order to solve the above technical problems, the present invention is realized through the following technical solutions:

[0011] In a first aspect, a cross-platform information comparison and consumer decision-making support system based on a mobile phone includes:

[0012] A multi-platform information retrieval module: supporting users to input keywords of goods or services through a unified search window and synchronously retrieving information from multiple network platforms;

[0013] A data integration and cleaning unit: preferentially obtaining real-time data through a platform interface, automatically switching to OCR parsing of user-authorization screenshots when the interface is unavailable, and processing data such as the names, prices, and promotion activities of goods / services based on a dynamic data standardization protocol (including BERT semantic matching and XGBoost false promotion identification);

[0014] Intelligent comparison and analysis module: Horizontally compare multi-platform data based on a preset algorithm to generate a visual comparison report;

[0015] Consumer decision-making support unit: Recommend the optimal option based on the comparison result and provide functions such as "one-click jump to place an order" or "integrated payment";

[0016] Cross-industry adaptation framework: Support modular expansion to multiple fields such as takeout, travel, and tourism, and each field can configure exclusive comparison dimensions.

[0017] In the first aspect, multi-dimensional information integration and intelligent decision-making support are achieved through the technical architecture. The core innovation lies in constructing a triangular interactive ecosystem of "consumer - physical store - platform". The specific technical details are as follows:

[0018] Multi-platform information retrieval module

[0019] Establish a dual-track mechanism of "interface first, OCR supplement". When the platform interface is unavailable, automatically switch to OCR parsing of the user-authorization screenshot. And only extract the key information of specific products during single collection, supporting millisecond-level synchronous acquisition of data from mainstream e-commerce, local life, and travel platforms. For example:

[0020] Obtain standardized data through official interfaces such as Meituan OpenAPI and Ele.me Business Connect. Users actively upload historical order screenshots (the system provides a standardized template)

[0021] Achieve "one-frame search, full-network response", with a response delay < 800ms

[0022] Data integration and cleaning unit

[0023] The dynamic data standardization protocol integrates BERT semantic matching and XGBoost false promotion recognition algorithm. When the data returned by the interface is abnormal, automatically trigger secondary verification to solve the problem of cross-platform data heterogeneity:

[0024] Naming normalization: Achieve semantic matching based on the BERT model (such as unifying "Hamburger Set A" and "Value-for-Money Hamburger Combo" as "Classic Beef Burger Set")

[0025] Price structuring: Establish a multi-dimensional price model to intelligently analyze the relationship between "marked price / discounted price after coupon / discounted price"

[0026] Evaluation sentiment analysis: Extract keywords through NLP (such as "slow delivery") to generate an emotion heat map

[0027] 2. Extended function design

[0028] Physical store data interface

[0029] Provide a SAAS management background and a standardized data access protocol:

[0030] Physical stores can upload exclusive coupon codes as a reference. Real-time data comparison still depends on the platform interface / OCR parsing (platform interface authentication is required).

[0031] Guarantee the authenticity of information through blockchain evidence preservation (e.g., a certain restaurant releases "50% off coupon exclusive to this tool").

[0032] Dynamic monitoring mechanism

[0033] Deploy a real-time data stream processing engine (Apache Flink) to achieve:

[0034] Price fluctuation warning: Trigger a reminder when the price returned by the authorized interface is 15% lower than the historical average price.

[0035] Rule parsing engine: Automatically mark restrictive conditions such as "no stacking of full reduction".

[0036] User behavior learning module

[0037] Optimize the recommendation strategy using privacy computing technology:

[0038] Federated learning framework: Train the user preference model on local devices (such as price sensitivity analysis).

[0039] Risk warning mechanism: Mark warning labels such as "there may be hidden charges".

[0040] 3. Compliance and security guarantee

[0041] Data acquisition protocol

[0042] Establish a three-layer compliance guarantee system:

[0043] Platform cooperation layer: Sign a data docking agreement with mainstream platforms (average daily API call volume ≤ 50,000 times).

[0044] User authorization layer: Need to check each item to access the platform (platforms without installed APPs are default closed).

[0045] Data processing layer: Retain raw data for ≤ 24 hours (only retain structured price comparison results).

[0046] Enhance payment security

[0047] Build a dual-channel fund supervision:

[0048] Platform direct connection payment: Jump to the official payment pages of each platform

[0049] System guarantee payment: Achieve fund splitting through the online clearing system (real-time fund splitting and custody in the supervision account of China Merchants Bank (account number: XXXXXX), and complete the settlement within 800ms after the payment instruction is generated).

[0050] Data emergency protocol: When both the interface and OCR fail, restricted web crawler technology (requiring a special legal agreement) can be used to obtain necessary price comparison data after secondary authorization by the user.

[0051] In a specific implementation of the first aspect, the physical store data interface: allows physical stores to directly upload product / service information and preferential policies, supplementing the platform data;

[0052] Dynamic monitoring mechanism: Real-time tracking of price fluctuations and promotion rule changes on each platform, and triggering intelligent reminders;

[0053] User behavior learning module: Optimize the recommendation algorithm based on historical operation records and provide personalized consumption suggestions.

[0054] In a specific implementation of the first aspect, the data collection protocol: preferentially obtain data through the platform interface, when the interface is unavailable, obtain screenshots of the mobile phone interface through the user's active authorization and use OCR technology for parsing, and limit the extraction of key information of specific products only for single collection;

[0055] Payment security guarantee: Integrate third-party payment channels or the platform's own payment interface to support cross-platform unified payment;

[0056] Data privacy protection module: Localize the user's screenshot data, encrypt and store sensitive information, and do not upload the original image to the server.

[0057] In a specific implementation of the first aspect, the market supervision assistance function: Generate statistical data such as industry price indices and platform service quality scores for reference by regulatory departments;

[0058] Merchant service optimization interface: Feedback user evaluations and suggestions to the platform to promote service quality improvement.

[0059] Second aspect, a method for cross-platform information comparison and consumption decision support based on the mobile phone side, including the following steps:

[0060] S1: The user inputs keywords of the target product / service;

[0061] S2: Within the scope of user authorization, preferentially obtain real-time data through the platform interface, when the interface is unavailable, collect product information from multiple platforms through screenshots, and perform OCR parsing and API data cleaning in parallel;

[0062] S3: Generate a multi-dimensional comparison report based on preset rules;

[0063] S4: Generate a recommendation index based on the federated learning framework and output the optimal consumption plan;

[0064] S5: Support the user to complete the order placement or payment within the system.

[0065] In the second aspect, the closed-loop of "information collection - analysis - decision - feedback" is achieved through process innovation, and its core value lies in reshaping the consumer decision-making logic. The specific implementation path is as follows:

[0066] 1. Detailed Explanation of the Core Process

[0067] S1: Keyword Input and Intention Analysis

[0068] Adopt semantic understanding technology (such as the BERT model) to identify the implicit needs input by users:

[0069] When the user searches for "Sichuan cuisine restaurants nearby", conditions such as "distance ≤ 3 km" and "per capita consumption of 50 - 80 yuan" are automatically extracted.

[0070] Privacy protection mechanism: Sensitive word filtering (such as not recording the user's mobile phone number and ID number entered).

[0071] S2: Authorized Data Aggregation and Standardization

[0072] Establish a workflow driven by user authorization:

[0073] Platform API docking: First, obtain real-time data through official interfaces such as Meituan / Ele.me (user authorization is required for each item). When the interface is unavailable, automatically switch to OCR parsing of the user's authorized screenshot.

[0074] User supplementary data: Guide the user to upload screenshots of historical orders (the system provides a standardized template) as a reference. Real-time data comparison does not rely on this.

[0075] Data cleaning:

[0076] Rule filtering: Exclude abnormal products with monthly sales < 10 orders (based on the public data returned by the interface).

[0077] Model verification: The dynamic data standardization protocol integrates the XGBoost algorithm to identify false promotions and works in coordination with BERT semantic matching (only analyzes the promotion information of the platforms authorized by the user).

[0078] S3: Generation of Multidimensional Comparison Reports

[0079] Visualization scheme optimization:

[0080] Dimensions shown in the radar chart: Price (interface data), rating (user authorized evaluation data), delivery speed (platform public data).

[0081] Thermodynamic table annotation rule: Only display the exclusive offers of the platforms with the APP installed by the user.

[0082] S4: Intelligent Recommendation and Risk Warning

[0083] Building an enhanced recommendation model:

[0084]

[0085] Calculation method of credit correction factor:

[0086] Based on the analysis of historical complaint data authorized by users (data retention ≤ 30 days).

[0087] The risk label only shows the change in the complaint rate of the platforms that the user has visited.

[0088] S5: Closed-loop payment and feedback collection

[0089] Payment security enhancement plan:

[0090] Payment jump: Directly evoke the official payment pages of each platform (without intercepting payment data).

[0091] Voucher deposit: Obtain the platform's electronic invoice after successful consumption, and deposit it in the blockchain of the regulatory account of China Merchants Bank after user confirmation.

[0092] Data feedback: User evaluation data is submitted through the official interface of the platform (secondary authorization is required).

[0093] In a specific implementation manner of the second aspect, the physical store can independently upload supplementary information as a reference basis, and real-time data comparison does not depend on this;

[0094] The system regularly generates a consumption trend analysis report for users' reference;

[0095] Intelligently warn against platform price discrimination and false promotion behaviors.

[0096] The beneficial effects of the present invention are as follows:

[0097] 1. The present invention preferentially obtains real-time data through the platform interface. When the interface is unavailable, it automatically switches to the OCR parsing of the user-authorization screenshot. Only the key information of specific products is extracted in a single collection, shortening the time-consuming of traditional multi-platform price comparison to 1.2 seconds. When both the interface and OCR fail, the restricted web crawler technology can be called with the user's secondary authorization to obtain the necessary price comparison data, realizing millisecond-level multi-platform information synchronous retrieval and shortening the time-consuming of traditional multi-platform price comparison from an average of 3.2 minutes to 1.2 seconds. The system uses BERT semantic parsing technology to identify the potential needs of users, combined with a dynamic weight recommendation model (recommendation index = price × 0.4 + score × 0.3 + timeliness × 0.2 + platform credit × 0.1), making the matching degree between the recommendation scheme and the user's actual choice reach 83%. The visual comparison report helps users quickly locate the optimal option, and together with the "platform direct connection for placing orders" and split-account payment functions, a complete consumption closed-loop is formed, shortening the decision-making time by 55% compared with the traditional method;

[0098] 2. Innovatively construct a triangular interactive ecosystem of "consumers - physical stores - platforms": Physical stores upload exclusive offers through the SAAS interface certified by the platform as a reference. Real-time data comparison still takes the interface / OCR parsing as the standard. The evaluation data shared with user authorization is only fed back through the official interface of the platform. The supplementary information of merchants is not used as a real-time data source. Combining blockchain evidence storage technology to break traffic monopolies; The dynamic monitoring mechanism parses the authorized interface data in real time (setting a 15% price reduction threshold warning) and implicit rules, and the accuracy rate of identifying price discrimination behavior reaches 92%. The industry price index generated by the system provides a decision-making basis for regulatory authorities. The evaluation data shared with user authorization is fed back through the official interface of the platform, promoting the overall improvement of the industry service quality by more than 20%. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] Figure 1 It is a schematic diagram of the system architecture of the present invention.

[0100] Figure 2 It is a detailed schematic diagram of the module composition of the present invention.

[0101] Figure 3 It is a schematic diagram of the data processing flow of the present invention.

[0102] Figure 4 It is a schematic diagram of user interaction of the present invention.

[0103] Figure 5 It is a schematic diagram of the deployment architecture of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0104] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0105] Such as Figures 1 to 5 A cross-platform information comparison and consumption decision support system and method based on the mobile phone side shown.

[0106] I. System Technical Architecture and Realization of Core Modules

[0107] 1. Distributed Data Collection and Processing Layer

[0108] Multi-source Data Access

[0109] First, obtain real-time data through the platform API interface (such as Meituan, Ele.me). When the interface is unavailable, automatically switch to OCR parsing of the screenshots authorized by the user, and only extract the key information of specific products (such as order history, coupons) for single collection.

[0110] Data emergency protocol: When both the interface and OCR fail, restricted web crawler technology can be invoked with secondary user authorization to obtain necessary price comparison data (a special legal agreement needs to be signed).

[0111] User-assisted collection: Provide a standardized screenshot template to guide users to upload multi-platform product page information as a reference basis. Real-time data comparison does not rely on this.

[0112] Data standardization engine

[0113] Named Entity Recognition (NER):

[0114] Train a BERT model to recognize product names (accuracy rate of 97.3%), such as unifying "KFC Golden Crispy Chicken Bucket" into "KFC Golden Crispy Chicken Set Meal".

[0115] Price structured processing:

[0116] Establish a price weight model to parse "Marked price is 59 yuan, 10 yuan off when reaching 50 yuan" into the actual payment price of 49 yuan.

[0117] 2. Intelligent Decision Analysis Layer

[0118] Dynamic configuration of comparison dimensions

[0119] Food delivery scenario: Price, full reduction rules, platform commission rate (obtain public data through the interface).

[0120] Travel scenario: Estimated price error rate (predicted based on the LSTM model, error rate ≤ 12%).

[0121] Implementation of recommendation algorithm

[0122] Basic model:

[0123] Score = (Price score × 0.4) + (Service score × 0.3) + (Timeliness score × 0.2) + (Platform credit score × 0.1).

[0124] Personalized optimization:

[0125] Use the multi-armed bandit algorithm to dynamically adjust the recommendation weights.

[0126] 3. User Interaction and Payment Layer

[0127] Unified payment gateway

[0128] Split payment: Funds are temporarily stored in the regulatory account of China Merchants Bank and are split according to the actual amount after consumption.

[0129] Payment interface: Directly jump to the official pages of WeChat Pay, Alipay, etc.

[0130] Intelligent Early Warning System

[0131] Abnormal Price Detection: Trigger an early warning when the price returned by the authorized interface is 15% lower than the historical average price.

[0132] False Propaganda Identification: After the user uploads the actual picture, perform image comparison locally (accuracy rate: 92%).

[0133] II. Industry Adaptation Solutions and Implementation Cases

[0134] 1. Adaptation for the Food Delivery Industry

[0135] Direct Connection of Physical Store Data: Upload coupons through the platform certification interface (order volume of a certain Hunan restaurant increased by 40%) as a reference, and real-time data comparison still relies on the interface / OCR parsing.

[0136] Commission Monitoring: Calculate the actual commission rates of Meituan / Ele.me (25% vs 22%).

[0137] 2. Application in the Travel Industry

[0138] Premium Index Analysis: The premium rate in a certain area during the morning rush hour is 35% (based on historical data authorized by users).

[0139] Driver Rating Screening: Prioritize displaying drivers with a score of ≥4.8.

[0140] 3. Solutions for the Tourism Industry

[0141] Hotel Monitoring: Push a reminder when the price of a certain hotel in Sanya subscribed by the user drops to 650 yuan.

[0142] Cancellation Policy Analysis: Use NLP to extract keywords such as "can be cancelled 7 days in advance".

[0143] III. Key Technology Verification Data

[0144] 1. Data Collection Efficiency

[0145] Response Time: 1.2 seconds (3 - 5 times faster than the traditional method).

[0146] Data Coverage: Support domestic platforms such as Meituan and Didi, as well as cross-border services such as Uber.

[0147] 2. Recommendation Accuracy

[0148] Matching Degree: 83%, decision-making time shortened by 55% (measured by one thousand users).

[0149] 3. Benefits of Physical Stores

[0150] Commission of Chain Fast Food Reduced by 18%, Monthly Order Volume Increased by 2000+.

[0151] Repurchase Rate of Small and Medium-sized Restaurants Increased to 35%.

[0152] Merchant supplementary information is only for display reference and is not included in the real-time price comparison data source.

[0153] IV. Implementation and Deployment Plan

[0154] 1. Cloud Architecture

[0155] Microservices architecture (Spring Cloud).

[0156] Database: MySQL + Elasticsearch + Redis.

[0157] 2. Mobile Development

[0158] Flutter framework to achieve cross-platform for iOS / Android (code reuse rate: 80%).

[0159] Weak network cache: Retain product data for 30 days.

[0160] In a weak network environment, interface cache data is preferentially used, and screenshot parsing is only triggered by the user's active price comparison operation.

[0161] Through technological innovation, cross-platform information transparency consumption has been achieved, and its implementation will reshape the market competition rules in the digital economy era. Based on the above refined plan, the system can be developed and piloted within 6 - 12 months, and is expected to cover the main cities across the country within 3 years, serving more than 100 million consumers, and promoting the improvement of the industry service quality by more than 20%.

[0162] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made in these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A cross-platform information comparison and consumption decision-making support system based on the mobile phone side, characterized in that: Including: Multi-platform information retrieval module: It supports users to input keywords of goods or services through a unified search window and retrieve information from multiple network platforms synchronously; Data integration and cleaning unit: It preferentially obtains real-time data through platform interfaces. When the interfaces are unavailable, it automatically switches to OCR parsing of user-authorization screenshots and processes data of goods / service names, prices, and promotional activities based on the dynamic data standardization protocol (including BERT semantic matching and XGBoost false promotion identification); Intelligent comparison and analysis module: It makes horizontal comparisons of multi-platform data based on preset algorithms and generates visual comparison reports; Consumer decision-making support unit: It recommends the optimal options according to the comparison results and provides functions of "one-click jump to place an order" or "integrated payment"; Cross-industry adaptation framework: It supports modular expansion to multiple fields such as takeout, travel, and tourism, and each field can configure exclusive comparison dimensions; 2. The cross-platform information comparison and consumption decision-making support system based on the mobile phone terminal according to claim 1, characterized in that: Physical store data interface: It allows physical stores to directly upload information of goods / services and preferential policies to supplement platform data; Dynamic monitoring mechanism: It tracks price fluctuations and changes in promotional rules of each platform in real time and triggers intelligent reminders; User behavior learning module: It optimizes the recommendation algorithm based on historical operation records and provides personalized consumption suggestions; 3. The cross-platform information comparison and consumption decision-making support system based on the mobile phone side according to claim 1, characterized in that: Data collection protocol: It preferentially obtains data through platform interfaces. When the interfaces are unavailable, it obtains screenshots of the mobile phone interface through user active authorization and uses OCR technology for parsing, and limits that only key information of specific goods is extracted in a single collection; Payment security guarantee: It integrates third-party payment channels or the platform's own payment interfaces and supports cross-platform unified payment; Data privacy protection module: It performs local processing on user screenshot data, encrypts and stores sensitive information, and does not upload the original image to the server; 4. A cross-platform information comparison and consumption decision support system based on a mobile phone terminal according to claim 1, characterized in that: Market supervision assistance function: It generates statistical data such as industry price indices and platform service quality scores for reference by regulatory departments; Merchant service optimization interface: It feedbacks user evaluations and suggestions to the platform to promote service quality improvement; 5. A method for cross-platform information comparison and consumption decision-making support based on a mobile phone terminal, characterized in that: Including the following steps: S1: The user inputs keywords of the target goods / services; S2: Within the scope of user authorization, it preferentially obtains real-time data through platform interfaces. When the interfaces are unavailable, it collects multi-platform goods information through screenshots and concurrently executes OCR parsing and API data cleaning; S3: Generates a multi-dimensional comparison report based on preset rules; S4: Generates a recommendation index based on the federated learning framework and outputs the optimal consumption plan; S5: Supports users to complete placing an order or payment within the system; 6. A method for cross-platform information comparison and consumption decision support based on a mobile phone terminal, characterized in that: Physical stores can independently upload supplementary information as a reference basis, and real-time data comparison does not depend on this; The system regularly generates a consumption trend analysis report for users' reference; Makes intelligent warnings about platform price discrimination and false promotion behaviors.

Images