Cost accounting method and system based on cost essential manifold
By constructing a multidimensional feature map and generating a probabilistic potential field based on the cost-essential manifold method, the problem of multidimensional feature variability and adaptability in DRG cost accounting is solved, and accurate dynamic cost calculation and optimization are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI XIRUAN TECH CO LTD
- Filing Date
- 2026-03-26
- Publication Date
- 2026-04-24
AI Technical Summary
Existing technologies cannot capture the multidimensional differences in the characteristics of medical service items in DRG cost accounting, ignore resource consumption, time sensitivity, and technical difficulty, resulting in biased accounting results and a lack of adaptability. They cannot respond to changes in the operating environment, which limits their application in decision-making such as performance evaluation and medical insurance negotiation.
A cost-based manifold approach is adopted, which constructs a cost gene map by acquiring multidimensional features, reduces the dimensionality to the essential manifold, generates a probabilistic potential field, and optimizes the parameters to achieve dynamic cost rate calculation.
It accurately reflects cost heterogeneity, adapts to changes in the operating environment, improves the accuracy and adaptability of accounting results, and supports departmental performance evaluation and treatment pathway optimization.
Smart Images
Figure CN121920794A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of cost accounting technology, specifically relating to a cost accounting method and system based on the cost manifold. Background Technology
[0002] In the context of current refined hospital operation and management, DRG (Diagnosis Related Groups) cost accounting has become a core component of medical resource management. However, existing technical systems have fundamental limitations that severely restrict their practical application value: traditional methods mainly rely on fixed cost rate formulas (such as a hospital-wide uniform "total cost / total revenue" ratio) or departmental accounting templates with limited hierarchical levels. Such static models cannot capture the inherent cost complexity of medical service items, ignore the differences in resource consumption, time sensitivity, technical difficulty, and other multidimensional characteristics of different items, and are difficult to respond to real-time changes in the operating environment (such as fluctuations in departmental load and resource inventory tightness), resulting in significant deviations between accounting results and actual costs. In particular, they cannot accurately reflect the cost heterogeneity of different departments within the same DRG group or due to different operating paths. A deeper problem is that existing methods lack adaptability and evolutionary capabilities. Once their parameters and rules are set, they tend to become rigid and cannot learn from historical data to optimize. When faced with dynamic scenarios such as updates to hospital service models and adjustments to resource structures, they appear inflexible and lagging, which greatly limits their supporting role in key decisions such as departmental performance evaluation, treatment pathway optimization, and medical insurance negotiations. Summary of the Invention
[0003] To address the aforementioned problems in the prior art, this invention provides a cost accounting method and system based on cost-essential manifolds.
[0004] The objective of this invention can be achieved through the following technical solutions: A cost accounting method based on cost-essential manifolds, the implementation of which includes the following steps: Step S1: Obtain the multidimensional features of each service item, and based on the multidimensional features, obtain the cost gene sequence of each service item and construct the project cost gene map; Step S2: Obtain the essential manifold coordinates of each service item based on the cost gene sequence, and reduce the dimensionality of the cost gene sequence to the cost essential manifold; Step S3: Obtain the probability density and generate the probability potential field based on the essential manifold coordinates; Step S4: Obtain the maximum a posteriori probability cost rate based on the probability potential energy field, and perform cost accounting based on the maximum a posteriori probability cost rate; Step S5: Establish a dual-scale optimization objective to optimize the probabilistic potential field parameters and the essential manifold coordinates.
[0005] Preferably, the construction of the project cost gene map in step S1 specifically involves: Extract the multidimensional features of each service item, including resource consumption genes, time-sensitive genes, risk-related genes, and departmental adaptation genes; The multidimensional features are encoded into standardized parameters to form the cost gene sequence; A complete mapping database from service items to cost gene sequences was established to obtain the cost gene map of the aforementioned items.
[0006] Preferably, step S2 specifically includes: Based on the cost gene sequence, label the N neighboring items with the highest similarity for each service item; The shortest path length between the service item and its neighboring items is obtained and denoted as the geodesic distance between the two items. The coordinates of the essential manifold are obtained through the geodesic distance, and the cost gene sequence is reduced to the cost essential manifold. Mathematically, this can be described as follows: ,in, For the essential manifold coordinates of service item i, To determine the geodetic distance between service project i and its neighboring project j, Let be the Euclidean distance between service item i and neighbor item j in the cost-essential manifold.
[0007] Preferably, the generation of the probabilistic potential field in step S3 specifically involves: S301: Obtain real-time operational data, which includes real-time load and resource inventory index, and use the real-time operational data as the field source; S302: Construct the load attractor function and inventory potential energy function based on the field source to obtain the attractor coordinates and inventory potential energy value; S303: Based on the attractor coordinates and the inventory potential value, construct a stochastic cost probability field function based on the field source on the cost essential manifold to generate the probability potential field.
[0008] Preferably, step S302 specifically includes: The load attractor function is constructed based on the real-time load, and mathematically described as follows: ,in, For attractor coordinates, Based on attractor coordinates, For element-wise multiplication, Where L is the load sensitivity coefficient and L is the real-time load. The benchmark load index; The inventory potential function is constructed based on the resource inventory index, and is mathematically described as follows: ,in, This represents the inventory potential value. Let r be the resource inventory index of the r-th resource. Let be the sensitivity coefficient of the r-th resource. Let r be the importance weight of the r-th resource. The degree of dependency of service item U on resource r.
[0009] Preferably, step S303 specifically includes: Based on the attractor coordinates and the stock potential energy value, an adjustment factor is obtained, mathematically described as follows: ,in, As a regulating factor, For load regulation sensitivity coefficient, For inventory adjustment sensitivity coefficient, The essential manifold coordinates of the service item U; The uncertainty factor, derived from the attractor coordinates and the inventory potential value, is mathematically described as follows: ,in, As an uncertainty factor, For load uncertainty sensitivity coefficient, This is the sensitivity coefficient to inventory uncertainty. Based on the adjustment factor and the uncertainty factor, the mean parameter and variance parameter are obtained, mathematically described as follows: ,in, The mean parameter, The base cost rate for service item U; ,in, For variance parameter, Based on the fundamental variance; The random cost probability field function is obtained based on the mean parameter and the variance parameter, and is mathematically described as follows: ,in, For a given The probability density of the cost rate CR under the given conditions. The mean is The variance is It follows a normal distribution.
[0010] Preferably, step S4 specifically includes: Based on the probabilistic potential field, a constrained optimization function is constructed on the cost-essential manifold to obtain the maximum a posteriori probability cost rate, mathematically described as follows: ,in, The maximum posterior probability cost rate. For safety margin.
[0011] Preferably, the dual-scale optimization objective in step S5 is specifically: ,in, To optimize the probabilistic potential field parameters, For optimized essential manifold coordinates, Let be the actual cost of the t-th sample. Let t be the fee for the t-th sample. Let be the maximum posterior probability cost rate for the t-th sample. For regularization strength, This is the manifold structure regularization term.
[0012] A cost accounting system based on cost-essential manifolds is used to execute the cost accounting method based on cost-essential manifolds described above, including a gene map construction module, a dimensionality reduction module, a probability potential field generation module, a cost accounting module, and an optimization module; The gene mapping module is used to obtain the multidimensional features of each service item, obtain the cost gene sequence of each service item based on the multidimensional features, and construct the project cost gene map. The dimensionality reduction module is used to obtain the essential manifold coordinates of each service item based on the cost gene sequence, and to reduce the dimensionality of the cost gene sequence to the cost essential manifold. The probability potential energy field generation module is used to obtain the probability density and generate the probability potential energy field based on the essential manifold coordinates. The cost accounting module is used to obtain the maximum posterior probability cost rate based on the probability potential energy field, and to perform cost accounting based on the maximum posterior probability cost rate. The optimization module is used to establish a dual-scale optimization objective and optimize the probabilistic potential field parameters and the essential manifold coordinates.
[0013] The beneficial effects of this invention are as follows: (1) By using manifold learning technology, the high-dimensional "cost gene" is compressed into a low-dimensional essential space, and a topological map that can deeply reflect the intrinsic cost similarity between projects is constructed, thus breaking through the coarseness of traditional classification methods.
[0014] (2) By introducing the theory of random probability fields, the cost rate is defined as a dynamic probability distribution modulated by multidimensional real-time environmental variables, rather than a single fixed value, thus realizing the leap from static calculation to dynamic situation perception of the accounting results.
[0015] (3) By simultaneously optimizing the representation coordinates of the project on the manifold and its corresponding cost behavior parameters through dual-scale learning, the system can accurately fit historical data and maintain the rationality and generalization ability of its topology during continuous learning and evolution. Attached Figure Description
[0016] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.
[0017] Figure 1 This is a flowchart illustrating the steps of a cost accounting method based on a cost-essential manifold according to the present invention. Detailed Implementation
[0018] To better understand the invention, various aspects of the invention will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are merely illustrative of exemplary embodiments of the invention and are not intended to limit the scope of the invention in any way. Throughout the specification, the expression "and / or" includes any and all combinations of one or more of the associated listed items. As used herein, the terms "approximately," "about," and similar terms are used as expressions of approximation, not as expressions of degree, and are intended to describe inherent deviations in measured or calculated values that will be recognized by those skilled in the art. Furthermore, the order in which the steps are described in this invention does not necessarily indicate the order in which these steps occur in actual operation, unless otherwise expressly defined or deduced from the context.
[0019] It should also be understood that expressions such as "comprising," "including," "having," "containing," and / or "comprising" are open-ended rather than closed-ended expressions in this specification, indicating the presence of the stated features, elements, and / or components, but not excluding the presence of one or more other features, elements, components, and / or combinations thereof. Furthermore, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features, not just individual elements in the list. Additionally, when describing embodiments of the invention, the word "may" is used to mean "one or more embodiments of the invention." And the term "exemplary" is intended to refer to examples or illustrations.
[0020] Unless otherwise specified, all terms used herein (including engineering and technical terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that, unless expressly stated herein, terms defined in common dictionaries shall be interpreted as having the meaning consistent with their meaning in the context of the relevant art, and not in an idealized or overly formalized sense.
[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other. The invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0022] Example 1: Please see Figure 1 A cost accounting method based on the cost-essential manifold includes: Step S1: Obtain the multidimensional features of each service item, and based on the multidimensional features, obtain the cost gene sequence of each service item and construct the project cost gene map; Step S2: Based on the cost gene sequence, obtain the essential manifold coordinates of each service item, and reduce the dimensionality of the cost gene sequence to a low-dimensional (generally 3-dimensional) cost essential manifold. For example, the coordinates of "laparoscopic cholecystectomy" are (0.12, -0.45, 0.78); the coordinates of "complete blood count" are (-0.30, 0.65, 0.10). Step S3: Obtain the probability density and generate the probability potential field based on the essential manifold coordinates; Step S4: Obtain the maximum posterior probability cost rate, i.e. the optimal and most reasonable cost rate, based on the probability potential energy field, and perform cost accounting based on the maximum posterior probability cost rate; Step S5: Establish a dual-scale optimization objective to optimize the probabilistic potential field parameters and the essential manifold coordinates.
[0023] In this embodiment, the construction of the project cost gene map is specifically as follows: S101: Extract the multidimensional features of each service item from the hospital's historical data. The multidimensional features include, but are not limited to, resource consumption genes (such as the intensity of consumable use, equipment dependence, and level of human resource input), time-sensitive genes (such as service duration standards, time period sensitivity, and seasonal fluctuations), risk-related genes (such as the probability of complications, technical difficulty coefficient, and quality sensitivity), and departmental adaptation genes (such as departmental specificity, cross-departmental applicability, and environmental dependence). S102: Encode the multidimensional features into standardized parameters to form the cost gene sequence; S103: Establish a complete mapping database from service items to cost gene sequences to obtain the cost gene map of the items.
[0024] In this embodiment, step S2 can be implemented through the following steps: S201: Based on the cost gene sequence, the N neighboring items with the highest similarity are labeled for each service item using the K-nearest neighbor algorithm; S202: The shortest path length between the service item and its neighboring items is obtained through Dijkstra's algorithm and denoted as the geodesic distance between the two items, which reflects the true difference in cost between the two items. S203: Obtain the coordinates of the essential manifold using the geodesic distance, and reduce the dimensionality of the cost gene sequence to the cost essential manifold, mathematically described as follows: ,in, For the essential manifold coordinates of service item i, Let the geodesic distance between service item i and neighbor item j be given in the original multidimensional space (X). Let be the Euclidean distance between service item i and neighbor item j in the cost-essential manifold (Y). This means finding a three-dimensional coordinate for each service item such that the straight-line distance between the service item and its neighbor item in the three-dimensional space (cost-essential manifold) is as close as possible to the geodesic distance in the multidimensional space.
[0025] In this embodiment, the generation of the probabilistic potential field is specifically as follows: S301: Obtain real-time operational data, which includes real-time load (such as operating room utilization rate) and resource inventory index (such as consumable resource inventory index, human resource inventory index, etc.), and use the real-time operational data as the field source; S302: Based on the field source, construct the load attractor function and the inventory potential energy function to obtain the attractor coordinates and the inventory potential energy value. The attractor coordinates move with the real-time load. When the load is high, the attractor coordinates will move to the region on the manifold that represents the high resource consumption project. The inventory potential energy value increases when the inventory of a certain resource is tight and the project depends on the resource, indicating that it is in a high potential energy region. S303: Based on the attractor coordinates and the inventory potential value, construct a stochastic cost probability field function based on the field source on the cost essential manifold to generate the probability potential field.
[0026] In this embodiment, step S302 can be implemented through the following steps: S302-1: Construct the load attractor function based on the real-time load, mathematically described as follows: ,in, For attractor coordinates, Based on attractor coordinates, For element-wise multiplication, Where L is the load sensitivity coefficient and L is the real-time load. The benchmark load index; S302-2: Construct the inventory potential function based on the resource inventory index, mathematically described as follows: ,in, This represents the inventory potential value. Let r be the resource inventory index of the r-th resource. Let be the sensitivity coefficient of the r-th resource. Let r be the importance weight of the r-th resource. The degree of dependency of service item U on resource r.
[0027] In this embodiment, step S303 can be implemented through the following steps: S303-1: An adjustment factor is obtained based on the attractor coordinates and the inventory potential energy value, mathematically described as follows: ,in, As a regulating factor, For load regulation sensitivity coefficient, For inventory adjustment sensitivity coefficient, The essential manifold coordinates of the service item U; S303-2: Based on the attractor coordinates and the inventory potential energy value, an uncertainty factor is obtained, mathematically described as follows: ,in, As an uncertainty factor, For load uncertainty sensitivity coefficient, This is the sensitivity coefficient to inventory uncertainty. S303-3: Based on the aforementioned adjustment factor and uncertainty factor, the mean parameter and variance parameter are obtained, mathematically described as follows: ,in, The mean parameter, The base cost rate for service item U; ,in, For variance parameter, Based on the fundamental variance; S303-4: The random cost probability field function is obtained based on the mean parameter and the variance parameter, mathematically described as follows: ,in, For a given The probability density of the cost rate CR under the given conditions. The mean is The variance is It follows a normal distribution.
[0028] In this embodiment, step S4 can be implemented through the following steps: Based on the probabilistic potential field, a constrained optimization function is constructed on the cost-essential manifold to obtain the maximum a posteriori probability cost rate, mathematically described as follows: ,in, The maximum posterior probability cost rate. A small safety margin (e.g., 0.05) is used; cost accounting is performed based on the maximum posterior probability cost rate. The cost of a service item is the sum of the costs of all services for a single patient, and so on.
[0029] Example: To obtain the cost of "laparoscopic cholecystectomy" in the current hospital setting: 1. Confirm that the essential manifold coordinates of this procedure are located at (0.12, -0.45, 0.78); 2. Obtain the real-time load L as 0.92 and obtain the attractor coordinates. ; Obtain the resource inventory index of 0.3 and the dependence of 0.8 to obtain the inventory potential value. 0.25; 3. Construct a stochastic cost probability field function and find the optimal one. The value is 0.74, and it is verified that it meets the constraints; 4. The project fee is 18,000 yuan, so its cost is 18,000 × 0.74 = 13,320 yuan.
[0030] In this embodiment, step S5 can be implemented through the following steps: Periodically compare actual costs; if the deviation exceeds a threshold, collect multiple samples within the deviation period and perform dual-scale optimization. The specific objective of the dual-scale optimization is... ,in, To optimize the probabilistic potential field parameters, For optimized essential manifold coordinates, Let be the actual cost of the t-th sample. Let t be the fee for the t-th sample. Let be the maximum posterior probability cost rate for the t-th sample. For regularization strength, For the regularization term of the manifold structure, the mathematical description is: .
[0031] Example 2: A cost accounting system based on a cost-essential manifold includes a gene map construction module, a dimensionality reduction module, a probabilistic potential field generation module, a cost accounting module, and an optimization module; The gene mapping module is used to obtain the multidimensional features of each service item, obtain the cost gene sequence of each service item based on the multidimensional features, and construct the project cost gene map. The dimensionality reduction module is used to obtain the essential manifold coordinates of each service item based on the cost gene sequence, and to reduce the dimensionality of the cost gene sequence to the cost essential manifold. The probability potential energy field generation module is used to obtain the probability density and generate the probability potential energy field based on the essential manifold coordinates. The cost accounting module is used to obtain the maximum posterior probability cost rate based on the probability potential energy field, and to perform cost accounting based on the maximum posterior probability cost rate. The optimization module is used to establish a dual-scale optimization objective and optimize the probabilistic potential field parameters and the essential manifold coordinates.
[0032] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.
Claims
1. A cost accounting method based on the cost-essential manifold, characterized in that, Includes the following steps: Step S1: Obtain the multidimensional features of each service item, and based on the multidimensional features, obtain the cost gene sequence of each service item and construct the project cost gene map; Step S2: Obtain the essential manifold coordinates of each service item based on the cost gene sequence, and reduce the dimensionality of the cost gene sequence to the cost essential manifold; Step S3: Obtain the probability density and generate the probability potential field based on the essential manifold coordinates; Step S4: Obtain the maximum a posteriori probability cost rate based on the probability potential energy field, and perform cost accounting based on the maximum a posteriori probability cost rate; Step S5: Establish a dual-scale optimization objective to optimize the probabilistic potential field parameters and the essential manifold coordinates.
2. The cost accounting method based on the cost-essential manifold according to claim 1, characterized in that, The construction of the project cost gene map in step S1 is specifically as follows: Extract the multidimensional features of each service item, including resource consumption genes, time-sensitive genes, risk-related genes, and departmental adaptation genes; The multidimensional features are encoded into standardized parameters to form the cost gene sequence; A complete mapping database from service items to cost gene sequences was established to obtain the cost gene map of the aforementioned items.
3. The cost accounting method based on the cost-essential manifold according to claim 1, characterized in that, Step S2 specifically includes: Based on the cost gene sequence, label the N neighboring items with the highest similarity for each service item; The shortest path length between the service item and its neighboring items is obtained and denoted as the geodesic distance between the two items. The coordinates of the essential manifold are obtained through the geodesic distance, and the cost gene sequence is reduced to the cost essential manifold. Mathematically, this can be described as follows: ,in, For the essential manifold coordinates of service item i, To determine the geodetic distance between service project i and its neighboring project j, Let be the Euclidean distance between service item i and neighbor item j in the cost-essential manifold.
4. The cost accounting method based on the cost-essential manifold according to claim 1, characterized in that, The generation of the probability potential field in step S3 is specifically as follows: S301: Obtain real-time operational data, which includes real-time load and resource inventory index, and use the real-time operational data as the field source; S302: Construct the load attractor function and inventory potential energy function based on the field source to obtain the attractor coordinates and inventory potential energy value; S303: Based on the attractor coordinates and the inventory potential value, construct a stochastic cost probability field function based on the field source on the cost essential manifold to generate the probability potential field.
5. The cost accounting method based on the cost-essential manifold according to claim 4, characterized in that, Step S302 specifically includes: The load attractor function is constructed based on the real-time load, and mathematically described as follows: ,in, For attractor coordinates, Based on attractor coordinates, For element-wise multiplication, Where L is the load sensitivity coefficient and L is the real-time load. The benchmark load index; The inventory potential function is constructed based on the resource inventory index, and is mathematically described as follows: ,in, This represents the inventory potential value. Let r be the resource inventory index of the r-th resource. Let be the sensitivity coefficient of the r-th resource. Let r be the importance weight of the r-th resource. The degree of dependency of service item U on resource r.
6. The cost accounting method based on the cost-essential manifold according to claim 5, characterized in that, Step S303 specifically includes: Based on the attractor coordinates and the stock potential energy value, an adjustment factor is obtained, mathematically described as follows: ,in, As a regulating factor, For load regulation sensitivity coefficient, For inventory adjustment sensitivity coefficient, The essential manifold coordinates of the service item U; The uncertainty factor, derived from the attractor coordinates and the inventory potential value, is mathematically described as follows: ,in, As an uncertainty factor, For load uncertainty sensitivity coefficient, This is the sensitivity coefficient to inventory uncertainty. Based on the adjustment factor and the uncertainty factor, the mean parameter and variance parameter are obtained, mathematically described as follows: ,in, The mean parameter, The base cost rate for service item U; ,in, For variance parameter, Based on the fundamental variance; The random cost probability field function is obtained based on the mean parameter and the variance parameter, and is mathematically described as follows: ,in, For a given The probability density of the cost rate CR under the given conditions. The mean is The variance is It follows a normal distribution.
7. The cost accounting method based on the cost-essential manifold according to claim 6, characterized in that, Step S4 specifically includes: Based on the probabilistic potential field, a constrained optimization function is constructed on the cost-essential manifold to obtain the maximum a posteriori probability cost rate, mathematically described as follows: ,in, The maximum posterior probability cost rate. For safety margin.
8. The cost accounting method based on the cost-essential manifold according to claim 1, characterized in that, The dual-scale optimization objective in step S5 is specifically... ,in, To optimize the probabilistic potential field parameters, For optimized essential manifold coordinates, Let be the actual cost of the t-th sample. Let t be the fee for the t-th sample. Let be the maximum posterior probability cost rate for the t-th sample. For regularization strength, This is the manifold structure regularization term.
9. A cost accounting system based on the cost-essential manifold, characterized in that, The system is applied to the cost accounting method based on cost-essential manifold as described in any one of claims 1-8, including a gene map construction module, a dimensionality reduction module, a probability potential field generation module, a cost accounting module, and an optimization module; The gene mapping module is used to obtain the multidimensional features of each service item, obtain the cost gene sequence of each service item based on the multidimensional features, and construct the project cost gene map. The dimensionality reduction module is used to obtain the essential manifold coordinates of each service item based on the cost gene sequence, and to reduce the dimensionality of the cost gene sequence to the cost essential manifold. The probability potential energy field generation module is used to obtain the probability density and generate the probability potential energy field based on the essential manifold coordinates. The cost accounting module is used to obtain the maximum posterior probability cost rate based on the probability potential energy field, and to perform cost accounting based on the maximum posterior probability cost rate. The optimization module is used to establish a dual-scale optimization objective and optimize the probabilistic potential field parameters and the essential manifold coordinates.