Commercial asset operation service system based on Internet of Things

Abstract

The invention provides a commercial asset operation service system based on the Internet of Things, which comprises a terminal sensing layer, an edge processing layer, a block chain layer and a cloud application layer, and the layers communicate through TLS1.3 encryption. The terminal sensing layer collects and encrypts original data, the edge processing layer decrypts and desensitizes the original data and then outputs desensitized data, the block chain layer realizes data storage through a dual hash and consensus mechanism and generates an audit log, and the cloud application layer provides a merchant management and third-party service interface. The system guarantees data privacy and non-tampering through edge desensitization-dual hash uplink-IPFS storage combination, and reduces on-chain pressure; fusing multi-dimensional data to construct a three-dimensional scoring model, predicting merchant potential based on LSTM and generating a revenue curve; the coupon is accurately pushed in combination with the user behavior and the merchant state, and the service efficiency and the cancel-after-verification rate are improved. According to the system, full-process safety processing and intelligent operation of data are realized, and prospective support is provided for commercial decision making.

Description

Technical Field

[0001] This invention relates to the field of commercial asset management, specifically a commercial asset operation service system based on the Internet of Things. Background Technology

[0002] Traditional commercial asset management systems are centered around centralized databases (such as MySQL and Oracle), relying on manual data entry or simple sensor data collection (such as water and electricity meter readings and personnel registration information) to achieve basic asset information archiving and retrieval. For example, a shopping mall management system enters information such as merchant rent and business hours through a PC-based backend, depends on manual inspections to record equipment status, has delayed data updates (usually once a day or once a week), and lacks real-time analysis capabilities.

[0003] With the development of the Internet of Things (IoT), IoT devices (such as RFID and temperature / humidity sensors) are being introduced into traditional commercial asset management systems to collect real-time data and transmit it to a cloud platform for centralized processing. For example, RFID tags are used to track inventory, smart POS machines record transaction data, and the cloud aggregates the data to generate sales reports. However, such systems experience high-frequency data uploads, leading to significant cloud pressure; furthermore, data storage relies on centralized servers, posing a risk of tampering.

[0004] Then, blockchain is used to store business data to ensure its immutability. For example, merchant transaction data is hashed and uploaded to a consortium blockchain to achieve traceability of transaction records. However, such solutions only focus on data storage and do not integrate real-time sensing data from IoT devices. They lack local processing and anonymization capabilities at the edge layer and do not involve operational analysis functions such as merchant profiling and potential prediction.

[0005] Meanwhile, existing merchant evaluations often rely on single-dimensional data (such as turnover). For example, a platform evaluates merchant performance based on monthly revenue rankings, which results in poor evaluation accuracy. Furthermore, preferential service strategies such as consumer vouchers are mostly based on fixed rules, lacking personalized adjustments based on user behavior, resulting in a lack of flexibility and poor adaptability.

[0006] The existing technology mainly has the following problems or defects:

[0007] I. Outdated data collection and processing modes: Traditional systems rely on manual input, resulting in poor real-time performance. Although basic IoT systems introduce device collection, they lack edge-layer local processing. High-frequency data (direct uploads cause cloud congestion). Furthermore, single blockchain solutions do not integrate real-time IoT data, and the stored evidence is limited to transaction records, which cannot support full-process operational decision-making.

[0008] Second, data security and credibility are insufficient. Both traditional systems and rudimentary IoT systems use centralized storage, which makes them prone to data tampering or leakage. In particular, rudimentary IoT systems have a higher risk of exposing sensitive data such as user payment information and merchant profit margins because the data is not desensitized. While a single blockchain solution can store evidence, it lacks edge-cloud collaboration and has low overall processing efficiency.

[0009] Third, there are limitations in merchant evaluation and service strategies. Existing evaluations rely heavily on a single revenue dimension, which cannot fully reflect core indicators such as business health and customer satisfaction. Dynamic service strategies are also subject to fixed designs and do not incorporate multi-dimensional data such as user dwell time and merchant potential, resulting in low accuracy. Furthermore, there is a lack of a linkage mechanism between multi-dimensional merchant evaluation and personalized services. Summary of the Invention

[0010] To address the aforementioned problems in existing technologies, this invention provides a business asset operation service system based on the Internet of Things (IoT).

[0011] A business asset operation service system based on the Internet of Things includes a terminal sensing layer, an edge processing layer, a blockchain layer, and a cloud application layer, with each layer connected via the TLS 1.3 encrypted communication protocol;

[0012] The terminal perception layer is used to collect raw data of commercial assets and encrypt it to generate encrypted data packets and signed data packets;

[0013] The edge processing layer is used to receive encrypted and signed data packets from the terminal perception layer, perform decryption, desensitization, local caching and real-time analysis, and output desensitized data;

[0014] The blockchain layer is used to receive de-identified data from the edge processing layer, realize data notarization through double hashing and consensus mechanism, and generate tamper-proof audit logs.

[0015] The cloud application layer is used to provide merchant management and third-party service interfaces based on anonymized data and tamper-proof audit logs, enabling data query, permission configuration, and personalized operation services.

[0016] Furthermore, the terminal sensing layer includes IoT devices and merchant terminals:

[0017] The IoT device is used to collect product inventory, transaction flow, and device data at a preset frequency. The original data is encrypted using the SM4 algorithm through a built-in encryption chip to generate an encrypted data packet containing device ID and timestamp. The device data includes environmental parameters and pedestrian flow data. The environmental parameters include temperature, humidity, light intensity, and air quality.

[0018] Merchants input business data into their terminals and sign the data using their private key to generate a signed data packet. The business data includes revenue and promotional strategies.

[0019] The encrypted data packets and the signed data packets together serve as input data for the edge processing layer.

[0020] Furthermore, the edge processing layer includes an edge computing gateway, a data desensitization unit, a local caching module, and a lightweight AI model;

[0021] The edge computing gateway is used to parse the encrypted data packets of the terminal perception layer via the MQTT protocol, and then send them to the data desensitization unit after decryption.

[0022] The data desensitization unit is used to perform differential processing on transaction flow, operational data and equipment data to generate desensitized data. The differential processing on transaction flow includes hiding the middle 6-12 digits of the card number. The differential processing on operational data includes range processing on the average order value and profit margin fields. The differential processing on equipment data includes extracting only the status identifier and not adding the original parameters to the blockchain.

[0023] The local cache module is used to temporarily store the original data before de-identification, for anomaly tracing, and is automatically destroyed after timeout;

[0024] The lightweight AI model is used to process local data in real time, including cleaning and format conversion of operational data, detecting outliers in environmental parameters, and outputting decision instructions.

[0025] The anonymized data serves as input data for the blockchain layer.

[0026] Furthermore, the blockchain layer includes consortium blockchain nodes, a smart contract module, and a hash-based evidence storage module:

[0027] The hash storage module is used to process the de-identified data output by the edge processing layer using SHA-256 + national cryptographic SM3 double hashing to generate a unique hash value. The original de-identified data is stored in the distributed file system IPFS, and only the IPFS address is retained on the chain.

[0028] The consortium blockchain nodes include verification nodes, storage nodes, and regulatory nodes. Verification nodes are responsible for verifying transactions; storage nodes are responsible for storing on-chain data; and regulatory nodes have read-only permissions. Verification nodes include platform providers and core merchants, while regulatory nodes include government agencies and industry associations.

[0029] The smart contract module is used to deploy permission management contracts, evidence storage contracts, and audit contracts. The generated audit logs and hash values ​​together serve as a trusted data source for the cloud application layer. The permission management contract includes defining roles and data access scopes, the evidence storage contract includes specifying the on-chain format, and the audit contract includes recording all data access operations and generating tamper-proof audit logs.

[0030] Furthermore, the cloud application layer includes a merchant management platform and third-party service interfaces:

[0031] The merchant management platform is used to perform data queries through private key signature verification, return the IPFS address and decryption key, and support permission configuration, including authorizing financial institutions to access transaction data for a specified period.

[0032] The third-party service interface is used to provide supply chain finance and advertising applications with open APIs. When the interface is called, it triggers the permission management contract verification and returns de-identified data or statistical results.

[0033] The merchant management platform also includes a 3D merchant profile building unit, a potential prediction unit, and a dynamic service unit, which realize multi-dimensional merchant evaluation and personalized services based on blockchain-layered evidence data.

[0034] Furthermore, the 3D merchant profile construction unit calculates the score using the following formula:

[0035] Business health score :

[0036] ;

[0037] In the formula, This represents the number of orders placed that day. This represents the park's average daily order volume. This is the daily turnover. This represents the park's average daily turnover. Product turnover rate, which is the ratio of sales volume to average inventory. The weighting coefficients are preferably 0.4, 0.4, and 0.2, respectively.

[0038] Customer satisfaction rating :

[0039] ;

[0040] In the formula, To evaluate the total number of items, For users to the first The rating of each product The sentiment weight of the comment text is calculated as follows: negative comments are weighted by 0.5, and positive comments are weighted by 1.2.

[0041] Resource utilization rate score :

[0042] ;

[0043] In the formula, For real-time pedestrian flow, In order to accommodate the flow of people, Daily energy consumption This serves as the industry's energy consumption benchmark.

[0044] Furthermore, the merchant development potential prediction unit uses LSTM to score merchant potential based on a three-dimensional rating, and the calculation formula is as follows:

[0045] ;

[0046] ;

[0047] In the formula, For the future Potential rating For the first The eigenvectors of the period, As a seasonally adjusted factor, for Weight matrix.

[0048] Furthermore, the dynamic service unit executes the following personalization strategy:

[0049] When the average order value per user exceeds the first preset threshold, coupons for high-end products with a face value of 10%-20% of the threshold will be automatically pushed to the user.

[0050] When a user stays for a longer period than the second preset threshold and does not place an order, the shopping guide robot's chat interface is triggered to provide product consultation and recommendation services.

[0051] Furthermore, the first preset threshold is between 100 yuan and 10,000 yuan, and the second preset threshold is between 5 minutes and 60 minutes.

[0052] Furthermore, the first preset threshold is 500 yuan, and the second preset threshold is 10 minutes.

[0053] The present invention has the following specific beneficial effects:

[0054] 1. Data security and efficient collaborative processing

[0055] By combining local desensitization at the edge processing layer, dual hashing on-chain, and IPFS storage, a dual guarantee of data privacy and immutability is achieved: local desensitization avoids the leakage of original sensitive data, dual hashing combined with the blockchain consensus mechanism ensures the authenticity and traceability of on-chain data, and IPFS distributed storage effectively reduces on-chain storage pressure and improves system operating efficiency; the four-layer collaborative design from the terminal perception layer to the cloud application layer constructs a closed loop for the entire process from data collection, encrypted transmission, local processing to intelligent services, ensuring the continuity and real-time nature of data processing.

[0056] 2. Accurate assessment and forward-looking prediction of merchant status

[0057] By integrating operational data, IoT device data, and sentiment analysis, a three-dimensional scoring model is built. Through a weighted algorithm, it breaks through the limitations of a single data dimension and achieves a comprehensive quantitative assessment of merchant status. Based on the three-dimensional profile input into the LSTM model, merchant potential is predicted, which can generate revenue curves for the next 6 months and support hypothesis analysis, providing data-driven forward-looking support for operational decisions and improving the accuracy and timeliness of decision-making.

[0058] 3. Personalized services and improved business efficiency

[0059] By combining user behavior data such as average order value and dwell time with the real-time status of merchants, dynamic and personalized service rules are constructed to achieve accurate coupon delivery and intelligent triggering of shopping guide scenarios, thereby improving service response efficiency. It accurately matches user needs with merchant discount strategies, effectively solving the problem of blind push in traditional methods, providing technical support for improving coupon redemption rates, and enhancing the actual conversion efficiency of business operations. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the structure of the Internet of Things-based commercial asset operation service system of the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] Please see Figure 1 This invention provides a business asset operation service system based on the Internet of Things, including a terminal sensing layer, an edge processing layer, a blockchain layer, and a cloud application layer, with each layer connected via the TLS 1.3 encrypted communication protocol.

[0063] The terminal perception layer consists of IoT devices deployed on-site at commercial assets and merchant terminals (PC / mobile management system).

[0064] The IoT devices collect raw data (product inventory, transaction flow, temperature and humidity, pedestrian flow, etc.) at a preset frequency (e.g., 10 seconds / time) and encrypt the data locally through the device's built-in encryption chip; the merchant terminal is used to enter and upload business data (e.g., daily revenue, promotional strategies).

[0065] IoT devices include, but are not limited to:

[0066] RFID tags: operating frequency 860-960MHz, identification distance 0-10 meters, used for collecting product inventory data;

[0067] Smart POS machines are used to collect transaction records, including fields such as transaction time, amount, and product details;

[0068] Environmental sensors: Temperature and humidity measurement ranges are -40℃ to 85℃ and 0 to 100%RH, respectively, used to collect temperature and humidity data;

[0069] Intelligent access control equipment: facial recognition and IC card recognition, through intelligent cameras and intelligent access control equipment to collect and record the time and movement path of people entering and exiting, for the purpose of collecting people flow lines;

[0070] Smart sensors: light sensor, measurement range 0-20000 lux; air quality sensor, can detect PM2.5, formaldehyde, etc., PM2.5 measurement range 0-500μg / m3; noise sensor, measurement range 30-130dB, used to collect light intensity, air quality and noise decibels in the store.

[0071] When in use, IoT devices collect raw data according to a preset cycle, encrypt it using the SM4 algorithm through a built-in encryption chip, and generate encrypted data packets (containing device ID, timestamp, and encrypted data). After the merchant terminal enters the business data, it signs the data with the merchant's private key to generate a signed data packet.

[0072] The edge processing layer consists of an industrial-grade edge computing gateway, a data desensitization unit, a local caching module, and a lightweight AI model.

[0073] The edge computing gateway receives encrypted data uploaded by the terminal perception layer, parses it using the MQTT protocol, and then sends it to the data desensitization unit.

[0074] The data desensitization unit performs differentiated processing for different data types, such as:

[0075] Transaction details: Hide the middle 6-12 digits of the card number, and keep the first and last 4 digits, such as 6226****1234;

[0076] Operational data: Fields such as average order value and profit margin are ranged, for example: average order value 500-1000 yuan;

[0077] Device data: Only status identifiers are extracted; raw parameters, such as normal or abnormal, are not uploaded to the blockchain.

[0078] The local cache module temporarily stores data before de-identification for anomaly tracing. It is automatically destroyed and retained after a timeout period of 24-72 hours.

[0079] In use, the edge computing gateway receives encrypted data packets, decrypts them with the device's public key, parses the data type (transaction / device / operational data), the data desensitization unit desensitizes the data according to preset rules, generates desensitized data, the ground cache module stores the original data (automatically deleted 24 hours later), and sends the desensitized data to the blockchain layer.

[0080] Lightweight AI models are used to process local data in real time, including preliminary processing such as cleaning and format conversion of operational data, and outlier detection of environmental parameters to output decision instructions, such as adjusting air conditioning temperature and lighting brightness.

[0081] The blockchain layer consists of consortium blockchain nodes, smart contract modules, and hash-based evidence storage modules.

[0082] The consortium blockchain nodes include verification nodes, storage nodes, and regulatory nodes. Verification nodes are composed of platform operators and core merchants and are responsible for transaction verification; storage nodes are responsible for on-chain data storage; and regulatory nodes have read-only permissions, such as government agencies and industry associations.

[0083] The smart contract module deploys three types of contracts:

[0084] Access control contract: Define roles (merchants, third parties, administrators) and data access scope (e.g., third parties can only access anonymized transaction statistics data);

[0085] The data storage contract specifies the format for uploading data to the blockchain and automatically triggers hash calculations. The data uploading format includes data ID, anonymized content, timestamp, and device identifier.

[0086] Audit contract: Records all data access operations and generates an immutable audit log.

[0087] The hash-based evidence storage module performs double hashing on the de-identified data using SHA-256 + national cryptographic SM3 to generate a unique hash value. Only the hash value and data index are uploaded to the blockchain, while the original de-identified data is stored in the distributed file system (IPFS), and only the IPFS address is retained on the blockchain.

[0088] When in use, the hash evidence storage module performs double hashing on the de-identified data, generating hash value H1 using the SHA-256 algorithm and hash value H2 using the national cryptographic SM3. H1+H2 is used as a unique identifier. The data ID, H1+H2, IPFS address, and timestamp are packaged into a transaction and sent to the verification node. The verification node uses the PBFT consensus mechanism to verify the legality of the transaction. After successful verification, it is written into the block. The permission management contract records the data ownership and accessible roles and generates an access control list.

[0089] The cloud application layer consists of a merchant management platform and third-party service interfaces.

[0090] The merchant management platform provides data query (requires private key signature verification) and permission configuration (e.g., authorizing a supply chain finance institution to access transaction statistics data for the past 3 months).

[0091] The third-party service interface is provided through an open API for applications such as supply chain finance and advertising. When called, it requires authorization management contract verification, and after verification, it returns anonymized data or statistical results.

[0092] When using the platform, merchants sign their requests with their private keys when querying data. After the access control contract verifies the request, it returns the IPFS address and decryption key. When a third party calls the API, the audit contract automatically records the caller ID, time, and data range, generates audit logs, and uploads them to the blockchain.

[0093] The merchant management platform also includes a 3D merchant profile building unit, a merchant development potential prediction unit, and a dynamic service unit.

[0094] The three-dimensional merchant profile building unit is used to conduct three-dimensional scoring of business health, customer satisfaction, and resource utilization.

[0095] The formula for calculating business health is:

[0096]

[0097] In the formula, To score the health of the business, This represents the number of orders placed that day. This represents the park's average daily order volume. This is the daily turnover. This represents the park's average daily turnover. Product turnover rate, which is the ratio of sales volume to average inventory. The weighting coefficients are preferably 0.4, 0.4, and 0.2, respectively.

[0098] The formula for calculating customer satisfaction is:

[0099]

[0100] In the formula, Rate customer satisfaction. To evaluate the total number of items, For users to the first The rating of each product The sentiment weight of the comment text is calculated as follows: negative comments are weighted by 0.5, and positive comments are weighted by 1.2.

[0101] The formula for calculating resource utilization rate is:

[0102]

[0103] In the formula, Score resource utilization rate For real-time pedestrian flow, In order to accommodate the flow of people, Daily energy consumption This serves as the industry's energy consumption benchmark.

[0104] The merchant development potential prediction unit uses LSTM to score merchant potential based on a three-dimensional rating, and the calculation formula is as follows:

[0105]

[0106]

[0107] In the formula, For the future Potential rating For the first The eigenvectors of the period, As a seasonally adjusted factor, for Weight matrix;

[0108] The execution logic of the dynamic service unit includes:

[0109] When the average order value per user exceeds the first preset threshold (which can be set by the merchant, ranging from 100 to 10,000 yuan, with a default value of 500 yuan), a high-end product coupon will be automatically pushed (with a face value of 10%-20% of the preset threshold and a validity period of 7 days).

[0110] When a user stays for a longer period than the second preset threshold (which can be set by the merchant, ranging from 5 to 60 minutes, with a default value of 10 minutes) and has not placed an order, the shopping guide robot conversation interface is triggered to provide services such as answering product inquiries and recommending products.

[0111] This invention has the following characteristics:

[0112] 1. By combining edge-layer local desensitization, dual hashing on-chain, and IPFS storage, we can ensure data privacy and immutability while reducing on-chain storage pressure. We adopt a collaborative design of terminal perception layer, edge processing layer, blockchain layer, and cloud application layer to achieve end-to-end processing from data collection to intelligent services.

[0113] 2. By integrating operational data, IoT device data, and sentiment analysis, a three-dimensional rating model is built. A weighted algorithm is used to comprehensively assess the merchant's status. Using the three-dimensional profile as input, LSTM is used to predict the merchant's potential, generating a revenue curve for the next 6 months and supporting hypothesis analysis, providing forward-looking support for operational decisions.

[0114] 3. By combining user behavior data such as average order value and dwell time with personalized service rules based on merchant status, we can achieve accurate coupon delivery and shopping guide triggering, improve service efficiency, and ensure the increase of coupon redemption rate.

[0115] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A business asset operation service system based on the Internet of Things, characterized in that: It includes a terminal perception layer, an edge processing layer, a blockchain layer, and a cloud application layer, with each layer connected via the TLS 1.3 encrypted communication protocol; The terminal perception layer is used to collect raw data of commercial assets and encrypt it to generate encrypted data packets and signed data packets; The edge processing layer is used to receive encrypted and signed data packets from the terminal perception layer, perform decryption, desensitization, local caching and real-time analysis, and output desensitized data; The blockchain layer is used to receive de-identified data from the edge processing layer, realize data notarization through double hashing and consensus mechanism, and generate tamper-proof audit logs. The cloud application layer is used to provide merchant management and third-party service interfaces based on anonymized data and tamper-proof audit logs, enabling data query, permission configuration, and personalized operation services.

2. The system as described in claim 1, characterized in that, The terminal sensing layer includes IoT devices and merchant terminals: The IoT device is used to collect product inventory, transaction flow, and device data at a preset frequency. The original data is encrypted using the SM4 algorithm through a built-in encryption chip to generate an encrypted data packet containing device ID and timestamp. The device data includes environmental parameters and pedestrian flow data. The environmental parameters include temperature, humidity, light intensity, and air quality. Merchants input business data into their terminals and sign the data using their private key to generate a signed data packet. The business data includes revenue and promotional strategies. The encrypted data packets and the signed data packets together serve as input data for the edge processing layer.

3. The system as described in claim 1, characterized in that, The edge processing layer includes an edge computing gateway, a data desensitization unit, a local caching module, and a lightweight AI model; The edge computing gateway is used to parse the encrypted data packets of the terminal perception layer via the MQTT protocol, and then send them to the data desensitization unit after decryption. The data desensitization unit is used to perform differential processing on transaction flow, operational data and equipment data to generate desensitized data. The differential processing on transaction flow includes hiding the middle 6-12 digits of the card number. The differential processing on operational data includes range processing on the average order value and profit margin fields. The differential processing on equipment data includes extracting only the status identifier and not adding the original parameters to the blockchain. The local cache module is used to temporarily store the original data before de-identification, for anomaly tracing, and is automatically destroyed after timeout; The lightweight AI model is used to process local data in real time, including cleaning and format conversion of operational data, detecting outliers in environmental parameters, and outputting decision instructions. The anonymized data serves as input data for the blockchain layer.

4. The system as described in claim 1, characterized in that, The blockchain layer includes consortium blockchain nodes, a smart contract module, and a hash-based evidence storage module. The hash storage module is used to process the de-identified data output by the edge processing layer using SHA-256 + national cryptographic SM3 double hashing to generate a unique hash value. The original de-identified data is stored in the distributed file system IPFS, and only the IPFS address is retained on the chain. The consortium blockchain nodes include verification nodes, storage nodes, and regulatory nodes. Verification nodes are responsible for transaction verification; storage nodes are responsible for storing on-chain data; and regulatory nodes have read-only permissions. Verification nodes include platform providers and core merchants, while regulatory nodes include government agencies and industry associations. The smart contract module is used to deploy permission management contracts, evidence storage contracts, and audit contracts. The generated audit logs and hash values ​​together serve as a trusted data source for the cloud application layer. The permission management contract includes defining roles and data access scopes, the evidence storage contract includes specifying the on-chain format, and the audit contract includes recording all data access operations and generating tamper-proof audit logs.

5. The system as described in claim 1, characterized in that, The cloud application layer includes a merchant management platform and third-party service interfaces: The merchant management platform is used to perform data queries through private key signature verification, return the IPFS address and decryption key, and support permission configuration, including authorizing financial institutions to access transaction data for a specified period. The third-party service interface is used to provide supply chain finance and advertising applications with open APIs. When the interface is called, it triggers the permission management contract verification and returns de-identified data or statistical results. The merchant management platform also includes a 3D merchant profile building unit, a potential prediction unit, and a dynamic service unit, which realize multi-dimensional merchant evaluation and personalized services based on blockchain-layered evidence data.

6. The system as described in claim 5, characterized in that, The 3D merchant profile building unit calculates the score using the following formula: Business health score : ； In the formula, This represents the number of orders placed that day. This represents the park's average daily order volume. This is the daily turnover. This represents the park's average daily turnover. Product turnover rate, which is the ratio of sales volume to average inventory. The weighting coefficients are preferably 0.4, 0.4, and 0.2, respectively. Customer satisfaction rating : ； In the formula, To evaluate the total number of items, For users to the first The rating of each product The sentiment weight of the comment text is calculated as follows: negative comments are weighted by 0.5, and positive comments are weighted by 1.

2. Resource utilization rate score : ； In the formula, For real-time pedestrian flow, In order to accommodate the flow of people, Daily energy consumption This serves as the industry's energy consumption benchmark.

7. The system as described in claim 5, characterized in that, The merchant development potential prediction unit uses LSTM to score merchant potential based on a three-dimensional rating, and the calculation formula is as follows: ； ； In the formula, For the future Potential rating For the first The eigenvectors of the period, As a seasonally adjusted factor, for Weight matrix.

8. The system as described in claim 5, characterized in that, The dynamic service unit executes the following personalized strategy: When the average order value per user exceeds the first preset threshold, coupons for high-end products with a face value of 10%-20% of the threshold will be automatically pushed to the user. When a user stays for a longer period than the second preset threshold and does not place an order, the shopping guide robot's chat interface is triggered to provide product consultation and recommendation services.

9. The system as described in claim 8, characterized in that, The first preset threshold is between 100 yuan and 10,000 yuan, and the second preset threshold is between 5 minutes and 60 minutes.

10. The system as described in claim 9, characterized in that, The first preset threshold is 500 yuan, and the second preset threshold is 10 minutes.

Images