CN-122022463-A - Intelligent enterprise cash flow prediction and risk early warning method and system

CN122022463ACN 122022463 ACN122022463 ACN 122022463ACN-122022463-A

Abstract

The invention provides an intelligent enterprise cash flow prediction and risk early warning method and system, which relate to the technical field of data processing, wherein the method comprises the steps of obtaining enterprise cash flow data; the method comprises the steps of constructing a cash flow stability tensor according to enterprise cash flow data, dividing the cash flow stability tensor into a promised cash flow tensor and an implementation cash flow tensor, determining a state depiction vector according to the promised cash flow tensor and the implementation cash flow tensor, constructing a hidden semi-Markov model according to the state depiction vector, carrying out state prediction on enterprise cash flow through the hidden semi-Markov model to obtain enterprise cash flow state distribution, constructing a risk infection map according to the enterprise cash flow state distribution, carrying out probability inference on the risk infection map to obtain risk probability of the enterprise cash flow, and carrying out risk early warning according to the risk probability.

Inventors

WANG YUXIANG
ZOU XIANG
CHEN JIANHUA
CHEN MINFENG
PU LIN

Assignees

无锡商业职业技术学院

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (10)

1. An enterprise cash flow intelligent prediction and risk early warning method is characterized by comprising the following steps: S1, acquiring cash flow data of an enterprise; s2, constructing a cash flow stability tensor according to the enterprise cash flow data; S3, dividing the cash flow stability tensor into a promised cash flow tensor and realizing the cash flow tensor; S4, determining a state depiction vector according to the promised cash flow tensor and the realized cash flow tensor; s5, constructing a hidden semi-Markov model according to the state describing vector; s6, carrying out state prediction on the cash flow of the enterprise through the hidden semi-Markov model to obtain state distribution of the cash flow of the enterprise; s7, constructing a risk infection chart according to the cash flow state distribution of the enterprise; S8, probability inference is carried out on the risk infection map to obtain the risk probability of the enterprise cash flow; And S9, performing risk early warning according to the risk probability.
2. The method for intelligent prediction and risk early warning of enterprise cash flow according to claim 1, wherein the step S2 specifically comprises: S201, preprocessing the enterprise cash flow data, including time granularity mapping, missing value processing and abnormal value suppression, to obtain a steady cash flow sequence group; S202, performing multi-scale consistency stability calculation on the steady cash flow sequence group to generate a stability tensor unit; And S203, performing multi-channel mapping and Zhang Lianghua assembly on the stability tensor unit to construct the cash flow stability tensor.
3. The method for intelligent prediction and risk early warning of enterprise cash flow according to claim 1, wherein the step S4 is specifically: S401, performing time granularity alignment and missing value complementation on the promised cash flow tensor and the realized cash flow tensor respectively to obtain an optimized promised cash flow tensor and an optimized realized cash flow tensor; S402, constructing a plurality of basic indexes according to the optimized promised cash flow tensor and the optimized realized cash flow tensor; And S403, splicing the basic indexes to obtain the state depiction vector.
4. The method for intelligent prediction and risk early warning of enterprise cash flow according to claim 1, wherein the step S5 specifically comprises: s501, constructing the state depiction vector into a cash flow observation sequence according to time sequence; s502, defining a correlation hidden variable and a significance parameter of the cash flow observation sequence; s503, constructing a two-channel observation generation model according to the correlation hidden variable and the significance parameter; S504, carrying out marginalization on the two-channel observation generation model to obtain final emission distribution; S505, constructing state transition probability and duration distribution of enterprise cash flow hidden states; s506, constructing the hidden semi-Markov model by combining the final emission distribution, the state transition probability and the duration distribution.
5. The method for intelligent prediction and risk early warning of enterprise cash flow according to claim 1, wherein the step S6 specifically comprises: s601, initializing model parameters of the hidden semi-Markov model, wherein the model parameters comprise initial state distribution, a state transition matrix, emission distribution parameters and duration distribution parameters; S602, setting an optimization target of the hidden semi-Markov model; S603, carrying out forward-backward inference on the hidden semi-Markov model to obtain a state posterior; s604, extracting features of the state posterior to obtain predicted features; S605, updating initialized model parameters according to the prediction characteristics and with the aim of minimizing the optimization target; s606, predicting the state of the enterprise cash flow according to the updated model parameters to obtain the state distribution of the enterprise cash flow.
6. The method for intelligent prediction and risk pre-warning of cash flow in enterprises according to claim 5, wherein S603 specifically comprises: s6031, calculating the observation probability of a cash flow observation sequence; s6032, carrying out forward inference on the hidden semi-Markov model according to the observation probability to obtain a forward probability; s6033, carrying out backward inference on the hidden semi-Markov model according to the observation probability to obtain a backward variable; s6034, calculating a segment posterior according to the forward probability and the backward variable; And S6035, calculating the state posterior according to the segment level posterior.
7. The method for intelligent prediction and risk early warning of enterprise cash flow according to claim 1, wherein the step S7 specifically comprises: S701, constructing an enterprise cash flow association network, wherein the enterprise cash flow association network comprises a node set and an edge set; S702, constructing a node potential function according to the node set and the enterprise cash flow state distribution; S703, defining an edge function according to the edge set; s704, combining the node potential function and the edge potential function to construct the risk infection map.
8. The method for intelligent prediction and risk early warning of enterprise cash flow according to claim 1, wherein the step S8 specifically comprises: s801, initializing message variables of the risk infection map to obtain an initial message set; S802, carrying out message updating on the initial message set according to a node potential function and an edge potential function; S803, judging whether the risk infection map has local risk concentration phenomenon or not, if so, entering a step S804, otherwise, determining the updated initial message set as a target message set, and entering a step S805; S804, performing diffusion suppression on the updated initial message set to obtain a target message set; s805, carrying out normalization processing on the target message set to obtain a normalized message; s806, calculating the risk edge probability of all nodes in the risk infection map according to the normalization message; s807, judging whether the maximum iteration number is reached, if so, entering a step S808, otherwise, returning to the step S802 until the maximum iteration number is reached; S808, combining the risk edge probabilities to obtain the risk probability of the enterprise cash flow.
9. An enterprise cash flow intelligent prediction and risk early warning system is characterized by comprising a processor and a memory; the memory stores a program or instructions executable on the processor, which when executed by the processor, implement the steps of the enterprise cash flow intelligent prediction and risk pre-warning method as claimed in any one of claims 1 to 8.
10. A readable storage medium having stored thereon a program or instructions which when executed by a processor performs the steps of the enterprise cash flow intelligent prediction and risk pre-warning method of any one of claims 1 to 8.

Description

Intelligent enterprise cash flow prediction and risk early warning method and system Technical Field The invention relates to the technical field of data processing, in particular to an intelligent enterprise cash flow prediction and risk early warning method and system. Background Enterprise cash flow is a core indicator reflecting enterprise business conditions and financial health levels, with operational status directly related to enterprise performance, capital turnover efficiency, and overall risk level. Through continuous monitoring, prediction and risk early warning of enterprise cash flows, potential fund shortage or performance risks can be found in time, enterprises are assisted to reasonably arrange fund plans, resource allocation is optimized, and coping capacity for operation uncertainty is improved. The process has important significance for enterprise financial management, risk control and robust operation, and is an important technical link for realizing enterprise fine financial management and risk prevention and control. Currently, for analysis and prediction of enterprise cash flow risk, a sequential time sequence modeling method based on historical cash flow data is generally adopted, for example, statistical analysis is performed on actual pay-and-receive data according to the occurrence time sequence of cash flow, and future cash flow conditions are predicted and early-warned through simple trend analysis, regression models or threshold judgment based on rules. The method is relatively simple in implementation mode, low in calculation cost and easy to deploy in the existing financial system, and can reflect the overall change trend of cash flow of enterprises to a certain extent. However, cash flow analysis methods based on sequential time series are typically modeled only on actual cash flow results, and it is difficult to simultaneously characterize the business's differential relationships between the planning level and the execution level, resulting in significant hysteresis in risk identification. Meanwhile, the method is mainly provided with the assumption that cash flow states frequently change between adjacent time points, the staged continuous characteristics of cash flow risks in actual operation cannot be effectively reflected, and under the conditions of complex business structure, dense fund association relation or frequent external environment changes, the risk evolution path and the potential systematic risks are difficult to accurately identify, so that the accuracy and the foresight of cash flow risk prediction and early warning are limited. Disclosure of Invention In view of the above-mentioned shortcomings of the prior art, an object of the present invention is to provide an enterprise cash flow intelligent prediction and risk early warning method, which can solve the technical problems that in the prior art, it is difficult to simultaneously describe the difference relationship between the plan level and the execution level of an enterprise, so that risk identification has obvious hysteresis, and the stepwise continuous characteristics of cash flow risks in actual operation cannot be effectively reflected, and it is difficult to accurately identify a risk evolution path and a potential systematic risk under the conditions of complex business structure, dense fund association relationship or frequent external environment changes, thereby limiting the accuracy and prospective technical problems of cash flow risk prediction and early warning. In a first aspect of the embodiment of the present invention, an intelligent enterprise cash flow prediction and risk early warning method is provided, including: S1, acquiring cash flow data of an enterprise; s2, constructing a cash flow stability tensor according to the enterprise cash flow data; S3, dividing the cash flow stability tensor into a promised cash flow tensor and realizing the cash flow tensor; S4, determining a state depiction vector according to the promised cash flow tensor and the realized cash flow tensor; s5, constructing a hidden semi-Markov model according to the state describing vector; s6, carrying out state prediction on the cash flow of the enterprise through the hidden semi-Markov model to obtain state distribution of the cash flow of the enterprise; s7, constructing a risk infection chart according to the cash flow state distribution of the enterprise; S8, probability inference is carried out on the risk infection map to obtain the risk probability of the enterprise cash flow; And S9, performing risk early warning according to the risk probability. In a second aspect of the embodiment of the invention, an intelligent enterprise cash flow prediction and risk early warning system is provided, which comprises a processor and a memory; the memory stores programs or instructions executable on the processor which when executed by the processor implement the steps of the enterprise cash flow intelligent pred