Bill arrangement system for business administration

Abstract

The invention discloses a bill arrangement system for business administration. The bill arrangement system comprises an electronic bill receiving module, an AI intelligent identification module, a block chain storage module, an intelligent contract module and a data query and analysis module. The system receives multi-format electronic bills, bill classification and information extraction are realized through an AI intelligent identification module by adopting a convolutional neural network and an attention mechanism, bill data are stored after Hash operation through a block chain storage module, and the security and non-tampering performance of the data are ensured. And the intelligent contract module verifies the bill classification confidence and the data integrity to ensure the authenticity of the data. The data query and analysis module provides bill query, statistics and anomaly detection functions, identifies money amount anomaly, and supports visual analysis and classified money amount summarization. According to the method, the automation level of electronic bill management is improved, the problems of low identification efficiency, insufficient data security and difficulty in anomaly detection are solved, and the method is suitable for the fields of enterprise financial management and business administration and has a wide application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of bill sorting, and particularly to a bill sorting system for business administration. Background Art

[0002] With the rapid development of the digital economy and the improvement of enterprise informatization, electronic bills have gradually replaced traditional paper bills and become important vouchers in business administration, enterprise financial management, and tax compliance. The promotion of electronic bills not only improves the speed and convenience of bill transmission but also plays an important role in saving paper, reducing enterprise costs, and minimizing manual operations. However, there are still many problems in the management, storage, recognition, and application of current electronic bills, which cannot meet the increasingly complex enterprise management requirements.

[0003] Currently, electronic bills mainly include forms such as invoices, receipts, and tax vouchers, and the circulation volume of these bills in daily business is huge. For example, the number of electronic bills processed by a medium-sized enterprise per day can reach hundreds, and the electronic bill flow of large enterprises is even in the thousands or tens of thousands. However, enterprises generally rely on traditional manual entry and simple classification methods for the management of electronic bills, resulting in low efficiency and a high error rate. The main problems include:

[0004] The traditional electronic bill management process relies on manual entry and classification, which is cumbersome, time-consuming, and error-prone. For example, financial personnel need to manually extract key information (such as amount, tax rate, date, etc.) on the bill and enter it into the financial system or tax declaration platform, which is time-consuming and laborious. For enterprises that process hundreds of bills every day, the error rate of manual entry and classification can be as high as 5% - 10%, seriously affecting the accuracy and efficiency of financial management.

[0005] The formats of electronic bills from different sources are not unified, and common formats include PDF, XML, and JSON, etc. The layouts, fonts, and structures of these bills are different, and traditional template matching methods and simple OCR (Optical Character Recognition) technologies are difficult to adapt to the complex and diverse bill formats. Especially when the bill contains complex elements such as tables, seals, and two-dimensional codes, the accuracy rate of traditional recognition technologies is relatively low and difficult to meet the actual needs.

[0006] Electronic bills are usually stored in a centralized database within the enterprise, posing risks of data loss, tampering, and illegal access. Especially during the data transmission process, the lack of a secure encryption mechanism makes it vulnerable to malicious attacks, resulting in data leakage. In addition, the centralized storage method cannot effectively ensure the immutability of data, affecting the authenticity and credibility of bill data.

[0007] During the process of financial auditing and tax inspection, enterprises need to trace the generation, transmission, and modification history of electronic invoices. However, existing systems lack an effective invoice traceability mechanism and it is difficult to accurately restore the invoice lifecycle. Insufficient data traceability leads to loopholes in the invoice verification process, increasing tax compliance risks.

[0008] Existing electronic invoice management systems mainly have basic data storage and query functions, lacking the ability to automatically detect and statistically analyze abnormal invoices. For example, enterprises cannot quickly identify issues such as abnormal amounts, duplicate invoices, or misclassification, which affects financial audits and risk management. In addition, the system lacks in-depth analysis and visualization functions for invoice data and is difficult to provide support for enterprises' financial decision-making.

[0009] Electronic invoices involve an enterprise's core financial data, and refined permission management is required for different role users (such as financial personnel, auditors, administrators). However, existing systems have deficiencies in access control and cannot achieve precise role-based authorization, easily leading to data leakage or misoperations. In addition, there is a lack of an effective encryption mechanism during the transmission of electronic invoices, making it difficult to guarantee data security.

[0010] Facing the above problems, electronic invoice management systems on the market mainly rely on traditional information technology and database storage methods and are difficult to meet enterprises' requirements for efficient, intelligent, and secure invoice management. The deficiencies of existing technologies are reflected in the following aspects:

[0011] Traditional OCR recognition methods have a low recognition rate when dealing with complex invoices (such as those with seals, handwritten content, or blurred images), with an accuracy of only about 80% - 85%.

[0012] For non-standard format invoices, template matching technology cannot be adapted, resulting in the system being unable to extract valid information.

[0013] Poor data security: Invoice data stored centrally is easily tampered with or lost, lacking the guarantee of data immutability.

[0014] No encryption measures are adopted during data transmission, making it vulnerable to man-in-the-middle attacks and posing a risk of leakage.

[0015] Lack of intelligent analysis and anomaly detection functions: Existing systems cannot automatically detect issues such as abnormal amounts and duplicate invoices, relying on manual checks, which is inefficient and prone to omissions.

[0016] Unable to classify and statistically analyze invoice data, summarize amounts, etc., and unable to meet the needs of enterprises for financial decision-making support.

[0017] Lack of system integration and scalability: The existing system has poor connectivity with the enterprise ERP and tax management systems, making data synchronization difficult. The system has poor scalability and is difficult to support the batch processing and management of a large number of electronic bills.

[0018] Therefore, we urgently need to design a bill sorting system for business administration to solve the above problems. Summary of the Invention

[0019] The purpose of the present invention is to provide a bill sorting system for business administration to solve the problems raised in the above background technology.

[0020] The above object of the present invention is achieved through the following solutions:

[0021] One aspect of the solution of the present invention provides a bill sorting system for business administration, including:

[0022] Electronic bill receiving module: used to receive electronic bill data. The input data X includes a format set F = {PDF, XML, JSON}, and the data size range D = [10KB, 10MB]. X: Input electronic bill data; F: Format set of electronic bills; D: Numerical range of bill data size;

[0023] A1 intelligent recognition module: performs feature extraction, attention weighting, and classification on the electronic bill image data X, and outputs a classification probability P(c|X), where c ∈ C = {c 1 , c 2 ,..., c k} is the bill category, and P(c|X): represents the probability that the input bill data X belongs to category c;

[0024] Blockchain storage module: used to generate a hash value H for the electronic bill data X, classification result c, and timestamp t, and record the data on the blockchain. H = SHA-256(T id ||X||c||t), where: T id is the unique identifier of the bill, X is the input bill data, c is the AI classification result, t records the timestamp, and the format is YYYY-MM-DD HH:MM:SS;

[0025] Smart contract module: used to record the generation, modification, classification, and verification status of bills to ensure the immutability and traceability of data;

[0026] Data query and analysis module: used to perform statistics, query, and anomaly detection based on the bill classification result c and the amount A storage record, where: A: The amount of the bill; c: The bill category result output by the AI recognition module.

[0027] Furthermore, the AI intelligent recognition module includes the following steps:

[0028] 1) Convolutional feature extraction is performed on the input bill X, and the calculation formula for the feature map F is:

[0029]

[0030] where: K = 3 is the convolutional kernel size, W m,n，k is the weight matrix of the k-th convolutional kernel, b k is the bias term, X i+m，j+n is the pixel value of the input bill image X at the position (i + m, j + n);

[0031] ReLU(x) = max(0, x) is the activation function;

[0032] 2) The convolutional feature map F is weighted through an attention mechanism to generate an attention distribution A:

[0033]

[0034] where: α is the attention coefficient, α ∈ [0.1, 2.0], Score(F i，j ) is the attention score function, A i，j is the attention weight at the (i, j) position;

[0035] 3) The output final classification probability P(c|X) is calculated by the following formula:

[0036] P(c|X) = Softmax(W c ·Attention(F) + b c ),

[0037] where: W c is the classification weight matrix, b c is the bias term, Attention(F) is the weighted feature map, c ∈ C is the bill category set, and P(c|X): represents the probability that the input bill data X belongs to the category c.

[0038] Furthermore, the blockchain storage module records the classification result c and the bill data X, and generates a verification hash value H n = SHA-256(H n-1 ||M n ||t n ), where: M n : the n-th modification content; t n : the current modification timestamp; H n-1 : the hash value of the previous record.

[0039] Further, the smart contract module verifies the confidence level P(c|X) of the bill classification result c and automatically triggers a successful verification when the following conditions are met:

[0040]

[0041] where: δ is the classification confidence threshold, and the value range is δ ∈ [0.7, 1.0],

[0042] V: verification result, V = 1 indicates passing, and V = 0 indicates failing.

[0043] Further, the data query and analysis module performs anomaly detection on the classification result c and the amount A, and filters the set of abnormal bills

[0044]

[0045] where: T i is the bill number, c i is the bill category, A i is the bill amount, τ is the abnormal amount threshold, and c 异常 is the preset abnormal category.

[0046] Further, the data query and analysis module further calculates the total amount S of the bills of each category c c :

[0047]

[0048] where: S c : the total amount of category c; δ(c i = c) is the category judgment indicator function. If c i = c, then δ = 1; otherwise, δ = 0; N is the total number of bills; A i is the amount of the i-th bill.

[0049] Further, the system supports role-based access control (RBAC), which is implemented through the mapping of user set U, role set R, and permission set P:

[0050] P = f(U, R), where R = {r 1 , r 2 , …, r n}.

[0051] Further, the system provides an API interface. Based on the encrypted transmission protocol TLS1.3, it synchronizes the bill data X, classification result c, and amount A to the enterprise ERP or financial management system to ensure transmission security, and the latency satisfies:

[0052] Δt ≤ 100ms.

[0053] Furthermore, the system supports distributed storage, and the response time of data nodes satisfies:

[0054] T n ≤ 10 ms, where n = the number of storage nodes.

[0055] Furthermore, the system provides a visualization analysis function, generating charts from the classification results c and the amount A to display the bill statistical results, the abnormal ratio, and the classification distribution.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] Through the AI intelligent recognition module of the present invention, combined with the convolutional neural network (CNN) and the attention mechanism, multi-format electronic bills (such as PDF, XML, JSON) are efficiently recognized and intelligently classified, solving the problems of low recognition accuracy and long manual entry time in traditional technologies. The system recognition accuracy can reach over 98%, significantly improving the automation level of electronic bill management and reducing the error rate caused by manual operations.

[0058] By adopting the blockchain storage module and the intelligent contract module, the key information of electronic bills (such as bill numbers, amounts, timestamps) is stored in a hash-encrypted manner to ensure the immutability and integrity of the data. In addition, the bill data is classified and verified through intelligent contracts to ensure the authenticity and credibility of the bill data, effectively solving problems such as data tampering and difficult traceability.

[0059] The system provides a data query and analysis module to achieve the automatic screening and intelligent detection of abnormal bills, and can quickly identify problems such as abnormal amounts, duplicate bills, or classification errors. At the same time, the system supports the amount aggregation and statistical analysis of various types of bills, and displays the results through visual reports to assist enterprises in making financial decisions, improving the enterprise's risk management and decision-making support capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0061] Figure 1 It is a system block diagram of a bill sorting system for business administration. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0062] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts fall within the scope of protection of the present invention.

[0063] The following describes in detail the specific implementation manners of the present invention in conjunction with multiple embodiments.

[0064] A bill sorting system for business administration includes:

[0065] An electronic bill receiving module: used to receive electronic bill data. The input data X includes a format set F = {PDF, XML, JSON}, and the data size range D = [10KB, 10MB]. X: the input electronic bill data; F: the format set of electronic bills; D: the numerical range of the bill data size;

[0066] An AI intelligent recognition module: extracts features, performs attention weighting, and classifies the electronic bill image data X, and outputs a classification probability P(c|X), where c ∈ C = {c 1 , c 2 ,..., c k} is the bill category, and P(c|X): represents the probability that the input bill data X belongs to category c. The AI intelligent recognition module includes the following steps:

[0067] 1) Perform convolutional feature extraction on the input bill X. The calculation formula for the feature map F is:

[0068]

[0069] where: K = 3 is the convolutional kernel size, W m,n，k is the weight matrix of the k-th convolutional kernel, b k is the bias term, and X i+m，j+n is the pixel value of the input bill image X at the position (i + m, j + n);

[0070] ReLU(x) = max(0, x) is the activation function;

[0071] 2) Weight the convolutional feature map F through the attention mechanism to generate an attention distribution A:

[0072]

[0073] where: α is the attention coefficient, α ∈ [0.1, 2.0], Score(F i，j ) is the attention scoring function, and A i，jis the attention weight at the (i, j) position;

[0074] 3) The final classification probability P(c|X) is calculated by the following formula:

[0075] P(c|X) = Softmax(W c ·Attention(F) + b c ),

[0076] where: W c is the classification weight matrix, b c is the bias term, Attention(F) is the weighted feature map, c ∈ C is the set of bill categories, and P(c|X): represents the probability that the input bill data X belongs to category c.

[0077] The blockchain storage module records the classification result c and the bill data X, and generates a verification hash value H n = SHA-256(H n-1 ||M n ||t n ), where: M n : the nth modified content; t n : the current modification timestamp; H n-1 : the hash value recorded last time.

[0078] The blockchain storage module: is used to generate a hash value H for the electronic bill data X, the classification result c, and the timestamp t, and record the data on the blockchain, H = SHA-256(T id ||X||c||t), where: T id is the unique identifier of the bill, X is the input bill data, c is the AI classification result, t records the timestamp, and the format is YYYY-MM-DD HH:MM:SS;

[0079] The smart contract module: is used to record the generation, modification, classification, and verification status of the bill, and ensure the immutability and traceability of the data; the smart contract module verifies the confidence level P(c|X) of the bill classification result c, and automatically triggers verification passing when the following conditions are met:

[0080]

[0081] where: δ is the classification confidence threshold, and the value range is δ ∈ [0.7, 1.0],

[0082] V: the verification result, V = 1 indicates passing, and V = 0 indicates not passing.

[0083] Data Query and Analysis Module: It is used to perform statistics, query, and anomaly detection based on the classification results c of bills and the stored records of amounts A, where: A: The amount of the bill; c: The bill category result output by the AI recognition module. The Data Query and Analysis Module performs anomaly detection on the classification results c and the amount A, and filters the set of abnormal bills

[0084]

[0085] Where: T i is the bill number, c i is the bill category, A i is the bill amount, τ is the abnormal amount threshold, c 异常 is the preset abnormal category.

[0086] The Data Query and Analysis Module further calculates the total sum S of the amounts of bills for each category c c :

[0087]

[0088] Where: S c : The total amount of category c; δ(c i =c) is the category judgment indicator function. If c i =c, then δ = 1, otherwise δ = 0; N is the total number of bills; A i is the amount of the i-th bill.

[0089] The system supports role-based access control (RBAC), which is implemented through the mapping of user set U, role set R, and permission set P:

[0090] P = f(U, R), where R = {r 1 , r 2 , …, r n}

[0091] The system provides an API interface. Based on the encrypted transmission protocol TLS1.3, it synchronizes the bill data X, classification results c, and amount A to the enterprise ERP or financial management system to ensure transmission security, and the latency meets:

[0092] Δt ≤ 100ms.

[0093] The system supports distributed storage, and the response time of data nodes meets:

[0094] T n ≤ 10ms, n = the number of storage nodes.

[0095] The system provides a visualization analysis function, generates charts from the classification results c and the amount A, and displays the bill statistical results, abnormal ratio, and classification distribution.

[0096] The following is a further elaboration in combination with several embodiments:

[0097] Embodiment 1: Overall system architecture and working process:

[0098] This system includes the following functional modules: Electronic bill receiving module; AI intelligent recognition module; Blockchain storage module; Smart contract module; Data query and analysis module;

[0099] The system architecture diagram is as Figure 1 shown, where the data source (payment platform, ERP system, tax system) provides bill data;

[0100] The electronic bill receiving module receives the data;

[0101] The AI intelligent recognition module extracts data features and classifies bill categories;

[0102] The blockchain storage module performs hash calculation on the data to ensure data security and immutability;

[0103] The smart contract module is responsible for data integrity verification and classification confidence judgment;

[0104] The data query and analysis module provides query, statistics, and abnormal bill detection functions.

[0105] Electronic bill receiving module: Input data source: The financial system of a certain enterprise, generating approximately 500 electronic bills every day, and the data format is F = {PDF, XML, JSON}.

[0106] Data size: The data size D of a single electronic bill ranges from 50 KB to 5 MB.

[0107] The data resolution is 1024×1024 pixels.

[0108] Receive electronic bill data from the data source and extract the bill number T id 、 amount A, and timestamp t.

[0109] Example data:

[0110]

[0111] The AI intelligent recognition module: Use a convolutional neural network (CNN) and an attention mechanism to perform feature extraction and classification on the input data X.

[0112] The convolutional kernel size K = 3, and the stride is 1.

[0113] The attention mechanism weights F and outputs the classification probability P(c|X):

[0114] P(c|X) = Softmax(Wc ·Attention(F)+b c ),

[0115] The value set of c: C = {invoice, receipt, tax voucher}.

[0116] Example classification result:

[0117]

[0118] Blockchain storage module: Generate a hash value H using the SHA-256 algorithm:

[0119] H = SHA-256(T id ||X||c||t),

[0120] Input: T id = 1001, c = invoice, t = 2024-06-15 09:00:00.

[0121] Generated hash value: H = fla2d9…5e6b7c.

[0122] Smart contract module: Verify the classification confidence P(c|X):

[0123]

[0124] If V = 0, mark the bill as abnormal data.

[0125] Data query and analysis module: Users can query bills by conditions and filter abnormal bills:

[0126]

[0127] Example: Set the abnormal amount threshold τ = 10,000 yuan, and the system filters out invoices with amounts exceeding the standard.

[0128] Example 2: Abnormal bill detection and classification statistics:

[0129] Abnormal detection process:

[0130] Input data:

[0131]

[0132] Filtering conditions: Abnormal amount threshold τ = 10,000. Abnormal category: invoice c 异常 = invoice.

[0133] System filtering result:

[0134]

[0135] Classification Statistics: Calculate the total amount S for different categories c :

[0136]

[0137] Total Invoice Amount: S 发票 = 15,000 + 5,000 = 20,000 yuan;

[0138] Total Receipt Amount: S 收据 = 12,000 yuan.

[0139] Example 3: RBAC Role - Based Access Control:

[0140] Role Setting: Administrator: Has access to all functions, including data query, statistics, storage, and modification.

[0141] Financial Staff: Only allowed to query bill data and statistical results.

[0142] Audit Staff: Only able to query abnormal bills.

[0143] Permission Mapping Relationship:

[0144] P = f(U, R), R = {Administrator, Financial Staff, Audit Staff}.

[0145] Example 4: System Performance and Security:

[0146] Data Encrypted Transmission: Use the TLS1.3 protocol for encrypted transmission to ensure data security. Data Transmission Delay:

[0147] Δt ≤ 100ms.

[0148] Distributed Storage Performance: The system adopts a distributed storage architecture, supporting dynamic expansion. Storage Node Response Time: T n ≤ 10ms, n = 20 nodes.

[0149] System Stability Test: Daily Bill Processing Volume: 10,000 bills; Abnormality Detection Time: 2s.

[0150] Finally, it should be noted that: The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: They can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A bill arrangement system for industrial and commercial management, characterized in that: include: Electronic bill receiving module: used to receive electronic bill data, input data X includes format set F = {PDF, XML, JSON}, data size range D = [10KB, 10MB], X: input electronic bill data; F: electronic bill format set; D: bill data size value range; AI intelligent recognition module: extract features, weight attention and classify the electronic receipt image data X, and output the classification probability P(c|X), where c∈C={c1,c2,…,c k } is the bill category, P(c|X): represents the probability that the input bill data X belongs to category c; Blockchain storage module: used to generate a hash value H for the electronic bill data X, classification result c and timestamp t, and record the data on the blockchain, H = SHA-256 (T id ||X||c||t), where: T id is the unique identifier of the bill, X is the input bill data, c is the AI ​​classification result, and t is the timestamp in the format of YYYY-MM-DD HH:MM:SS; Smart contract module: used to record the generation, modification, classification and verification status of bills to ensure the immutability and traceability of data; Data query and analysis module: used to perform statistics, query and anomaly detection based on the bill classification result c and amount A storage records, where: A: the amount of the bill; c: the bill category result output by the AI ​​recognition module.

2. A bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The AI ​​intelligent recognition module includes the following steps: 1) Perform convolution feature extraction on the input bill X. The calculation formula of the feature map F is: Where: K = 3 is the convolution kernel size, W m,n,k is the weight matrix of the kth convolution kernel, b k is the bias term, X i+m,j+n is the pixel value of the input bill image X at position (i+m,j+n); ReLU(x)=max(0,x) is the activation function; 2) The convolution feature map F is weighted through the attention mechanism to generate the attention distribution A: Where: α is the attention coefficient, α∈[0.1,2.0], Score(F i,j ) is the attention score function, A i,j is the attention weight of the (i, j)th position; 3) The final classification probability P(c|X) is calculated by the following formula: P(c|X)=Softmax(W c ·Attention(F)+b c ), Where: W c is the classification weight matrix, b c is the bias term, Attention(F) is the weighted feature map, c∈C is the bill category set, P(c|X): represents the probability that the input bill data X belongs to category c.

3. A bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The blockchain storage module records the classification result c and the bill data X, and generates a verification hash value H n =SHA-256(H n-1 ||M n ||t n ), where: M n : The nth modification content; t n : Current modification timestamp; H n-1 : The hash value of the last record.

4. A bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The smart contract module verifies the confidence level P(c|X) of the bill classification result c, and automatically triggers verification when the following conditions are met: Where: δ is the classification confidence threshold, the value range is δ∈[0.7,1.0], V: verification result, V=1 means passed, V=0 means failed.

5. The bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The data query and analysis module performs anomaly detection on the classification result c and the amount A, and filters the abnormal bill set Where: T i is the bill number, c i For the note type, A i is the bill amount, τ is the abnormal amount threshold, c 异常 The default exception category.

6. A bill arrangement system for industrial and commercial management according to claim 5, characterized in that: The data query and analysis module further calculates the total amount of bills S of each category c c : Where: S c : The total amount of category c; δ(c i =c) is the category judgment indicator function, if c i =c, then δ = 1, otherwise δ = 0; N is the total number of bills; A i is the amount of the i-th note.

7. A bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The system supports RBAC role-based access control, which is achieved through mapping of user set U, role set R and permission set P: P=f(U,R), where R={r1,r2,…,r n }.

8. The bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The system provides an API interface, based on the encrypted transmission protocol TLS1.3, to synchronize the bill data X, classification result c and amount A to the enterprise ERP or financial management system to ensure transmission security and delay satisfaction: Δt≤100ms.

9. A bill arrangement system for industrial and commercial management according to claim 8, characterized in that: The system supports distributed storage, and the data node response time meets: T n ≤10ms, n = number of storage nodes.

10. The bill arrangement system for industrial and commercial management according to claim 1, characterized in that: The system provides a visual analysis function, which generates a chart based on the classification result c and the amount A to display the bill statistics, abnormal proportions and classification distribution.

Images