9 results about "Tax administration" patented technology

The invention discloses an automatic settlement path generation method fusing a knowledge graph and a tax rule, and relates to the technical field of financial science and technology and intelligent finance and tax compliance, and the method comprises the steps: S1, obtaining initial decision basis data and execution process details from a settlement path through a block chain network, employing a Hash chain structure to link each link record, s2, according to the obtained path chain, a graph database is adopted to store dynamic association attributes, and the dependency relationship of each node in the path is determined through knowledge graph reasoning, and S3, if an information fault is detected in the path chain, the matching degree of an execution result and a decision basis is compared through a timestamp verification mechanism, and if the information fault is detected in the path chain, the decision basis is determined. Judging a deviation position and marking a deviation point; according to the automatic settlement path generation method fusing the knowledge graph and the tax rules, the security and reliability of financial transactions and the tax management efficiency are effectively improved.

The disclosure relates to the field of making online payments by integrating social networks and payment networks to a seamless network. In particular, some embodiments of the invention leverage the information in the social network to conduct financial transactions. A social payment network app can incorporate all or most of the functions of credit cards, internet banking and money order companies by using the data and functions already available in the social network. Some embodiments of the invention provide better counterparty recognition, reduce the chance of fraud, or fraud, and provide a cheaper payment platform to consumers or small businesses while also reducing the administrative burden in making and paying purchases or reporting payments and / or expenses to third parties, such as the employer or tax administration.

The invention discloses a digital electricity tax digital management system based on artificial intelligence, relates to the technical field of data management, and solves the problem that organic fusion of unified collection and intelligent risk identification of multi-source heterogeneous bills cannot be realized in the prior art. Through an innovative five-layer cloud native architecture, a data layer, a processing layer, an analysis layer, an application layer and a supporting layer are organically integrated, and a complete closed loop from bill collection to intelligent decision making is formed. The data layer realizes unified collection of bills, the processing layer completes data standardization management through a distributed calculation engine, the analysis layer constructs a three-dimensional enterprise portrait and a dynamic risk scoring model, the application layer provides full-life-cycle automatic management, intelligent early warning and one-key declaration of the bills, and the support layer guarantees safety compliance and elastic operation and maintenance of the system. According to the framework, a business finance and tax data chain is broken through, a risk identification gateway is moved forward, and the intelligent level and compliance risk control capability of tax management are remarkably improved.

The invention relates to the technical field of tax management, and discloses an enterprise tax risk control decision analysis method based on AI and big data processing technology, and the method comprises the steps: collecting the multi-source initial data of an enterprise, and carrying out the preprocessing of the multi-source initial data, and obtaining the preprocessed data; extracting static features, time sequence features and associated features from the preprocessed data to obtain multi-dimensional features; distributing attention weights to the multi-dimensional features, carrying out feature fusion through feature splicing and convolution operation, and outputting a fused high-dimensional feature vector; through an XGBoost model, a lightweight CNN (Convolutional Neural Network) model and a Transform model in a risk identification model, respectively carrying out risk preliminary identification on the high-dimensional feature vector, and carrying out adaptive weight adjustment on an output result of each model by adopting a meta-learning algorithm; based on a risk identification result output by the risk identification model, adopting a Markov decision process model to generate an optimized risk coping strategy; according to the invention, the risk identification accuracy is improved.

The invention discloses an enterprise taxpayer tax risk assessment method based on a knowledge graph. The method comprises the following steps: constructing a tax risk knowledge graph database based on a tax management system database; calculating an operating index anomaly threshold value of the enterprise taxpayer; calculating the operating index abnormality of each enterprise taxpayer; calculating the tax payment inmaturity and the tax payment trust level of the enterprise taxpayer; and calculating the weighted operation abnormity degree of each enterprise taxpayer and dividing the risk level of the enterprise taxpayer. According to the method, the knowledge graph, clustering analysis, anomaly detection and an entropy weight method are combined, the problems that in the prior art, data utilization is not comprehensive, industry scale differences are ignored and the like are solved, a multi-dimensional feature model containing operation indexes and behavior indexes, risk related feature indexes and a risk measurement algorithm are constructed, and an automatic risk assessment process is achieved; the tax risk assessment method can be applied to tax risk assessment scenes of various industries, and has wide application and popularization values.

This invention relates to the field of tax-related data processing technology and discloses a comprehensive tax administration platform for tax-related data collection and analysis. This system addresses the problem of how to break down data barriers between multiple departments and establish a unique data system centered on the real estate unit code in real estate tax management scenarios, a problem inherent in traditional methods. The system registers the sources, unifies fields, and standardizes time for tax-related data from multiple departments. It establishes a unique index for each property based on the real estate unit code, and interfaces to handle tax-related information throughout the entire lifecycle of a property according to its construction, filing, registration, transaction, leasing, ownership, and declaration relationships. The system performs traceability, risk management, pricing support, pre-filling collaboration, and declaration coordination around the property object. It also manages resource scheduling, access control, audit records, anomaly handling, and archiving control during the system's processing. Finally, it enables unified aggregation, traceability, correlation analysis, and collaborative processing of real estate tax-related management scenarios.

The invention discloses a test method for a new power bureau system, relates to the field of tax administration, and provides the test method for the new power bureau system in order to solve the problem that at present, test implementation for the new power bureau system still depends on manpower. And organizing a plurality of different tax data by taking the taxpayer ID as a unique identifier. And two business modes, namely a single-region single-business mode and a single-region multi-business mode, are distinguished according to the attribution and the declaration type for organization. And the file is output in a corresponding directory structure, so that the file is not mixed during storage, and a single-region single-service file or a single-region multi-service file can be conveniently read from a corresponding directory based on test requirements during subsequent pressure test. Therefore, the method can replace manual work to complete organization of tax data, is suitable for characteristic data in a tax scene, eliminates manual operation errors, can generate data files based on two different organization modes of single-region single business and single-region multi-business, and meets more diversified test requirements.

The invention relates to the technical field of tax management, in particular to an intelligent tax management method and system based on big data analysis. Comprising the steps of setting a plurality of business points based on business requirements, and constructing a tax review model according to all the business points; generating a plurality of tax feature structures based on a preset tax database, and setting a tax processing model according to all tax feature structures; and obtaining the sub-tax packages of each business point, and generating a first-level tax strategy according to all the sub-tax packages and the tax processing model. The business structure of the enterprise is analyzed, multiple business points are set, and the corresponding review sub-model is constructed according to the tax data category of each business point, so that the review early warning efficiency of tax risks in the enterprise is improved. According to the method, multiple tax structures are constructed, and the tax processing model is constructed according to the feature parameters of each tax structure, so that the linkage analysis efficiency and accounting declaration efficiency of tax data in an enterprise are improved, and the tax processing capability of the enterprise is improved.

The invention provides a tax processing method based on large language model AI intelligent question and answer, which comprises the following steps: enabling a model to have professional semantic understanding ability in the tax field through model fine tuning or prompt word engineering by utilizing a basic LLM (Logistics Language Model); collecting laws and regulations, policy interpretation, practical operation guide and case text data, extracting knowledge information through NLP structured processing, and integrating and constructing a tax knowledge base; combining an NLP processing flow, and fusing the extracted knowledge to construct a tax domain knowledge graph; performing fine adjustment on the basic LLM through the tax field corpus; aiming at complex tax problems, through thinking chain cue word design, a model is guided to disassemble the problems step by step, a conclusion is deduced in combination with laws and regulations in a knowledge base, and the logicality and compliance of answers are ensured; an API protocol is defined, and the LLM is in butt joint with an enterprise internal tax management system, a nationwide value-added tax invoice checking platform API and an electronic tax bureau declaration interface. According to the invention, the efficiency and compliance of enterprise tax management are improved.