An online price evaluation method and system based on big data ship depreciation rate curve fitting

Abstract

The present application relates to a kind of online price evaluation method and system based on big data ship depreciation rate curve fitting, method includes: obtaining multidimensional ship information, using multidimensional ship information to carry out multi-dimension correction to transaction price, obtain the integrated price under standard condition, i.e.standard price;Aggregation same ship type and the standard price corresponding to the standard price of age section is calculated benchmark point base, and with the minimum fitting objective function as target, through big data regression model fitting continuous benchmark value curve;Based on benchmark value curve, the market value of target ship is comprehensively evaluated, and the final evaluation price is obtained.The present application can be fitted by the depreciation rate curve of various parameters of ship, online ship, realize the dynamic, accurate and intelligent evaluation of ship asset value, suitable for ship valuation, second-hand ship transaction, insurance pricing, financing lease and asset management and multiple application scenarios.

Description

Technical Field

[0001] This invention relates to the field of intelligent shipping and asset valuation technology, and in particular to an online price valuation method and system based on big data-driven ship depreciation rate curve fitting. Background Technology

[0002] As important fixed assets, ships play a crucial role in valuation in secondhand ship transactions, financial leasing, insurance pricing, and corporate asset management. Existing ship valuation methods mainly include the market comparison approach, the cost approach, and the income approach. The market comparison approach uses recent transaction prices of similar ships as a basis for valuation; however, it is affected by factors such as significant individual differences between ships, low market transparency, and sparse transaction samples, leading to significant fluctuations in valuation results and insufficient reliability. The cost approach calculates present value based on the cost of new shipbuilding, combined with the ship's age and depreciation rate. However, traditional depreciation models typically use linear or fixed-rate depreciation, failing to accurately reflect the non-linear decline in ship value with age, technological advancements, and policy changes, especially the accelerated decline in value when the ship's age exceeds a critical point. The income approach values ​​ships based on future operating income; however, it is greatly affected by external factors such as cyclical fluctuations in the shipping market, fuel prices, and route rates, exhibiting strong subjectivity, poor stability, and difficulty in providing repeatable and reliable valuation results.

[0003] Furthermore, existing assessment systems largely rely on human experience and static rules, lacking the ability to systematically integrate and dynamically correct multi-source data. The multidimensional data involved in ship transactions, including historical transaction records, broker quotes, ship technical parameters, classification society reputation, construction location information, and navigation area characteristics, are difficult to fully quantify and utilize using traditional methods. This results in assessment results lagging behind market changes, making it impossible to achieve real-time, accurate, and interpretable ship valuation.

[0004] In summary, the existing technology has the following prominent problems: (1) Insufficient data utilization: Failure to effectively integrate multi-source, multi-dimensional, and time-series ship data resulted in incomplete valuation information. Traditional methods typically rely on a single data source, such as historical transaction prices or new shipbuilding costs, combined with static depreciation models for valuation. This approach struggles to accurately reflect the non-linear value changes of vessels as they age, technology evolves, and the shipping market environment evolves. When vessels exceed a certain age or when the market experiences structural fluctuations, the valuation results often deviate from actual transaction prices, reducing their reference value and decision-making reliability.

[0005] (2) Poor model adaptability: Traditional depreciation models cannot fit the nonlinear decay law of ship market value and are difficult to reflect real market dynamics. Ship valuation is influenced not only by structured parameters such as ship type, deadweight tonnage, and year of construction, but also by non-quantifiable factors such as place of construction, classification society level, and navigation area characteristics. Traditional depreciation models and manual experience rules are insufficient to effectively quantify and correct for these multi-dimensional factors, resulting in valuation results that lack specificity and flexibility, and are unable to meet the accurate valuation needs of different types of ships.

[0006] (3) Lack of correction mechanism: There is a lack of quantitative correction means for unstructured characteristics such as classification society grade, place of construction, and navigation area, making it impossible to achieve comprehensive evaluation.

[0007] (4) Weak dynamic update capability: Existing assessment systems are unable to automatically optimize parameters and update results in response to market fluctuations, thus failing to meet the needs for real-time and accurate valuation. Existing valuation systems largely rely on manual experience or static rules for calculations, lacking dynamic learning and closed-loop update mechanisms. Faced with cyclical fluctuations in the shipping market, changes in fuel prices, and policy adjustments, valuation results cannot be updated in real time, causing ship asset valuations to lag behind market changes and hindering rapid decision-making and risk management.

[0008] (5) Insufficient real-time performance: Existing assessment systems mostly rely on manual experience or static rules for calculation, lacking dynamic learning and closed-loop update mechanisms. Faced with cyclical fluctuations in the shipping market, changes in fuel prices, and policy adjustments, the valuation results cannot be updated in real time, causing the valuation of ship assets to lag behind market changes and making it difficult to support rapid decision-making and risk management. Summary of the Invention

[0009] To address the problems existing in the prior art, the present invention aims to provide an online price assessment method and system based on big data-driven ship depreciation rate curve fitting. This method addresses several technical issues in existing ship valuation techniques, particularly in terms of assessment accuracy, model adaptability, real-time performance, and interpretability. By fitting various ship parameters online to the ship's depreciation rate curve, it achieves dynamic, accurate, and intelligent assessment of ship asset value. This method is applicable to various scenarios such as ship valuation, secondhand ship transactions, insurance pricing, financial leasing, and asset management.

[0010] To achieve the above objectives, the present invention provides the following solution: An online price assessment method based on big data-driven ship depreciation rate curve fitting includes: Step 1: Obtain multi-dimensional ship information, and use the multi-dimensional ship information to make multiple corrections to the transaction price to obtain a comprehensive price under standard conditions, i.e., the standard price; Step 2: Aggregate the standard prices corresponding to the same ship type and age range to calculate the benchmark base, and fit a continuous benchmark curve by using a big data regression model with the goal of minimizing the fitting objective function. Step 3: Based on the benchmark curve, conduct a comprehensive assessment of the market value of the target vessel to obtain the final assessed price.

[0011] Optionally, the multi-dimensional ship information includes: ship transaction price data, broker quotation data, basic ship information, and market macro indicators.

[0012] Optionally, obtaining the standard price includes: ; in, For standard price, This is the actual transaction price. For deadweight tons, It is a function of depreciation rate. For classification society correction factors, For the construction site correction factor, This is the correction factor for the flight area.

[0013] Optionally, the depreciation rate function includes: ; in, , For depreciation rate parameters, Year of construction The constants defined for the fuzzy expert method.

[0014] Optionally, the base for calculating the standard price corresponding to the same ship type and age range includes: in, For the same ship type Same ship age The base number of the reference point, For standard price, As weight, For the sample set, , representing each sample.

[0015] Optionally, the objective function for fitting is defined as follows: ; in, To fit the objective function, As the base point, For smoothing parameters, As a baseline function, for, This is the differential symbol.

[0016] Optionally, fitting a continuous baseline curve using the big data regression model includes: ; in, For B-spline basis functions, For learnable parameters, The total number of B-spline basis functions used. Number the basis functions.

[0017] Optionally, obtaining the final appraised price includes: ; in, For the target vessel's deadweight tonnage; The age of the target vessel; These are the latest correction coefficients for the target ship obtained through expert fuzzy reasoning; This is the base unit value for the corresponding ship age.

[0018] Optionally, the method further includes: A closed-loop update mechanism is used to update multi-dimensional ship information in real time, and steps 1-2 are repeated until a new baseline curve is obtained.

[0019] To achieve the above objectives, the present invention also provides an online price evaluation system based on big data ship depreciation rate curve fitting, comprising: The data acquisition module is used to obtain multi-dimensional ship information; The comprehensive price inverse calculation module is used to perform multiple corrections on the transaction price using the multi-dimensional ship information to obtain the comprehensive price under standard conditions, i.e., the standard price. The benchmark generation module is used to aggregate the benchmark base for calculating standard prices corresponding to the same ship type and age range. The benchmark fitting module is used to fit a continuous benchmark curve using a big data regression model with the goal of minimizing the fitting objective function. The assessment calculation module is used to comprehensively assess the market value of the target vessel based on the benchmark curve and obtain the final assessment price.

[0020] The beneficial effects of this invention are as follows: This invention, by constructing a hierarchical modeling architecture, can fully integrate multi-source data such as historical transaction data, broker quotes, ship technical parameters, classification society reputation, construction site information, and macro market indicators, achieving quantitative integration of multi-dimensional information. Simultaneously, this invention introduces an expert fuzzy reasoning mechanism to dynamically correct unstructured and fuzzy features, combined with nonlinear depreciation rate curve fitting technology, to achieve accurate calculation of ship asset value.

[0021] This invention not only dynamically responds to market changes and achieves real-time online updates, but also possesses excellent interpretability. Through a hierarchical model and data fusion strategy, the contribution of each data point and feature to the final valuation result can be clearly traced, thereby improving the transparency and credibility of the assessment. In summary, this invention, through a big data-driven, AI-assisted, and hierarchical modeling approach, effectively solves the technical problems of low accuracy, poor adaptability, insufficient real-time performance, and weak interpretability in existing ship valuation technologies. It achieves dynamic, accurate, and interpretable online valuation of ship assets, providing scientific and reliable technical support for secondhand ship transactions, financial leasing, insurance pricing, and corporate asset management.

[0022] This invention achieves several significant technical effects in ship asset valuation by introducing a hierarchical modeling architecture, expert fuzzy algorithms, a dynamic closed-loop update mechanism, and multi-source data fusion technology. Specifically, this invention solves the problems of low valuation accuracy, poor model adaptability, insufficient real-time performance, and weak interpretability in existing technologies, and brings significant social and economic benefits.

[0023] This invention significantly improves the accuracy of ship valuation. By constructing a comprehensive price inverse calculation and multi-maintenance correction mechanism, this invention can uniformly process data from different sources, achieving a refined fit to ship prices. The benchmark value fitting module employs spline regression and regularization techniques to smooth and continuously map discrete transaction samples, enabling the model to more accurately capture the nonlinear laws governing changes in ship value with age and market conditions. Experimental simulations and historical data validation show that the method of this invention outperforms traditional static depreciation models in valuation results for various ship types and navigation areas, significantly reducing valuation bias while maintaining high consistency and stability.

[0024] This invention enhances the adaptability and robustness of the model. Through an expert fuzzy algorithm, it can quantify and correct unstructured features, including classification society reputation, shipbuilding technology, and navigation area characteristics, enabling the model to maintain stable performance across different market cycles, ship types, and operating environments. Simultaneously, the time decay weight mechanism makes the model more sensitive to recent market data, thereby improving its response speed and adaptability to market fluctuations. Empirical analysis shows that this feature significantly improves the model's stability and prediction accuracy in complex and volatile market environments.

[0025] This invention features real-time dynamic updating capabilities. Through a closed-loop learning mechanism, new transaction data triggers partial model updates, achieving dynamic synchronization of evaluation results. The application of an incremental learning strategy avoids full retraining, improving system efficiency while ensuring the timeliness and continuity of valuation results. Simulated market change environments validate the invention's ability to maintain accuracy and rapid response even with frequent data updates.

[0026] This invention enhances the interpretability and transparency of the valuation process. Each correction coefficient is supported by a membership function and a rule base, enabling a complete decision traceability path. The output includes correction factor contribution analysis and confidence interval information, allowing experts and regulatory agencies to clearly track the evaluation process, facilitating review, auditing, and risk control, and significantly improving the credibility of the valuation results.

[0027] Furthermore, this invention brings significant economic and social benefits. By improving valuation accuracy, real-time performance, and interpretability, it provides a reliable basis for ship transactions, financing assessments, and insurance pricing, helping to optimize investment decisions and reduce transaction risks and financing costs. Simultaneously, this invention enhances the digitalization and intelligence of ship asset management, promotes the healthy development of shipping finance and the secondhand ship market, and has promising prospects for widespread application and social benefits.

[0028] In summary, through the synergistic effect of various features in the technical solution, this invention inevitably produces technical effects such as improving assessment accuracy, enhancing model adaptability and robustness, achieving real-time dynamic updates, and improving interpretability. At the same time, it brings economic and social value, providing a scientific, reliable, and widely applicable ship valuation method for shipping asset management. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram of an online price evaluation method based on big data ship depreciation rate curve fitting, according to an embodiment of the present invention. Detailed Implementation

[0031] 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, and 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.

[0032] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] like Figure 1As shown in the figure, this embodiment discloses an online price evaluation method based on big data ship depreciation rate curve fitting, including: Step 1, acquiring multi-dimensional ship information, using the multi-dimensional ship information to perform multi-correction on the transaction price, and obtaining a comprehensive price under standard conditions, i.e., a standard price; Step 2, aggregating the standard prices corresponding to the same ship type and age range to calculate the benchmark base, and fitting a continuous benchmark curve through a big data regression model with the goal of minimizing the fitting objective function; Step 3, comprehensively evaluating the market value of the target ship based on the benchmark curve to obtain the final evaluation price.

[0034] Specifically, this embodiment discloses an online price evaluation method based on big data ship depreciation rate curve fitting, including: Input data includes: ship transaction price data It is derived from historical transaction records and is a numerical time-series data that fluctuates with the market. Broker quote data The data is derived from monthly market price reports and is a numerical time-series data that fluctuates with the market. Basic ship information, including ship type (Categorical variable), deadweight tonnage (Numerical variables), construction site (Categorical variables), classification society (Categorical variable), Year of construction (Numerical variables), operating airspace (Categorical variables), some of which are updated over time (e.g., flight area adjustments); Depreciation rate function A continuous function derived from market statistics, updated over time; correction factors for classification society, place of construction, and navigation area. It is estimated by expert fuzzy algorithms based on the characteristics of the classification society, the place of construction and the navigation area, and belongs to the category of numerical and nonlinear fuzzy variables.

[0035] Output data includes: comprehensive price The price is a standardized price, which is numerical and updated over time; the base point is the benchmark. For a certain ship type Ship age range Average standardized value; benchmark function , indicating ship type Continuous age-value function; final valuation price , which represents the output valuation result.

[0036] The key variable types and attributes are shown in Table 1: Table 1 Receive actual transaction price and broker quotes By combining the depreciation rate function and expert correction factors, the comprehensive price under standard conditions is calculated. The calculation logic is as follows: First, calculate the standardized transaction ratio: ; Then, multiple maintenance corrections are performed based on the depreciation rate and correction factor: ; in, The depreciation attenuation factor is represented by a differentiable and continuous exponential-piecewise fitting form, which takes into account both long-term attenuation and short-term stability. ; in, It is a constant defined by the fuzzy expert method. This is the depreciation rate parameter. This form guarantees 0– Depreciation is relatively flat during the first year, then accelerates, which is consistent with the common depreciation pattern in the shipbuilding market. These represent the correction factors related to the classification society, place of construction, and navigation area, generated by the expert fuzzy algorithm. The calculation formula is as follows: ; in, Let fuzzy membership function be defined in the expert experience space. .

[0037] Specifically, regarding the space of expert experience Define the fuzzy interval using the expert method, and then apply it to the current... Membership degree calculation is performed, and the membership degree function is as follows: ; Where parameters The initial price is determined by an expert panel and can be fitted based on data. This module outputs a standardized composite price. This is used for subsequent benchmark point calculations.

[0038] For the same ship type Same ship age A baseline base is generated by clustering and weighting all standardized prices within the range. .

[0039] Define the sample set: ; The formula for calculating the base number of the benchmark is: ; Among them, weight Defined by the sample time decay function: ; in, This indicates the interval between the sample and the current time. This represents the time decay coefficient. The output is the discrete reference point base matrix corresponding to each ship type-age combination. .

[0040] Based on the base matrix of the reference point A continuous benchmark function is obtained by fitting a big data regression model. This enables a smooth mapping from discrete reference points to continuous curves.

[0041] Define the fitting objective function: ; The second term is the smoothing regularization term. For smoothing parameters.

[0042] Fitting the model using spline regression: ; in, For B-spline basis functions, These are learnable parameters.

[0043] The training objective is to minimize : ; The optimization uses batch gradient descent: ; The output is a smooth, continuous age-benchmark function. .

[0044] The market value of the target vessel is comprehensively assessed based on its fundamental parameters, including deadweight tonnage and age, and incorporating the latest correction coefficients obtained through expert fuzzy reasoning. The final valuation formula is as follows: ; in, For the target vessel's deadweight tonnage; The age of the target vessel; These are the latest correction coefficients for the target ship obtained through expert fuzzy reasoning; This serves as the benchmark unit value for the corresponding ship age. It is automatically adjusted based on updated data when the market experiences significant fluctuations. The coefficients ensure that the evaluation results are timely and robust.

[0045] Furthermore, the method also includes: using a closed-loop update mechanism to update multi-dimensional ship information in real time, and repeating steps 1-2 until a new baseline curve is obtained.

[0046] Specifically, the input data is updated in real time during operation. Whenever a new transaction record is recorded... The collected data is recalculated by the comprehensive price inversion module. This triggers local refitting in the benchmark point generation module.

[0047] In actual operation, it possesses real-time closed-loop update capabilities, dynamically responding to changes in new data. When a new transaction record... After being collected, the corresponding standardized comprehensive price is first recalculated by the comprehensive price back-calculation module. In this process, the depreciation rate function, expert fuzzy correction coefficients, and various local environmental parameters are considered simultaneously to ensure that the quantitative correction of the new data is accurate and stable.

[0048] Subsequently, the standardized new price will The benchmark point generation module is triggered to perform local refitting. This refitting process not only updates local cluster centers or reference price ranges but also fine-tunes the original correction coefficients and depreciation parameters based on the new data, thereby maintaining the model's sensitivity and adaptability to market changes. During this process, new price data and historical data are weighted to ensure that local updates do not have an unstable impact on the overall model, while preserving long-term trend information from historical data.

[0049] Through the aforementioned closed-loop mechanism, the model can develop adaptive learning capabilities during operation, achieving dynamic correction and continuous optimization. Regardless of the magnitude of market price fluctuations, it can respond quickly and update the standardized composite price and related benchmarks in real time. This ensures that subsequent cluster analysis, price forecasting, and risk assessment modules maintain high accuracy and reliability, thus forming a complete real-time closed-loop operating architecture that provides continuous, stable, and traceable data support for ship valuation and market analysis.

[0050] This invention is driven by real transaction data and adopts a structure combining expert fuzzy correction and big data regression modeling. Its innovations include: (1) establishing a standardized mechanism for comprehensive price inversion and multi-level correction; (2) realizing the transformation from discrete samples to continuous value curves through benchmark aggregation and continuous fitting; (3) introducing expert fuzzy algorithms to achieve quantitative correction of unstructured features; and (4) designing an adaptive update mechanism based on time-series regression to achieve dynamic market synchronization. The final output valuation price is... It has high accuracy and interpretability in areas such as ship asset transactions, financing assessments, insurance pricing, and market monitoring.

[0051] This method is highly interpretable: for each sample Each rule has a clearly defined generation path (membership vector + rule constant), facilitating the provision of an auditable decision chain during implementation. For newly emerging classification societies or construction sites, adaptation can be achieved quickly by expanding the rule base and using incremental training with small samples.

[0052] This embodiment also provides an online price evaluation system based on big data ship depreciation rate curve fitting, including: a data acquisition module for acquiring multi-dimensional ship information; a comprehensive price back-calculation module for using multi-dimensional ship information to perform multiple corrections on the transaction price and obtain a comprehensive price under standard conditions, i.e., a standard price; a benchmark point generation module for aggregating standard prices corresponding to the same ship type and age range to calculate the benchmark point base; a benchmark value fitting module for fitting a continuous benchmark value curve through a big data regression model with the goal of minimizing the fitting objective function; and an evaluation calculation module for comprehensively evaluating the market value of the target ship based on the benchmark value curve and obtaining the final evaluation price.

[0053] Specifically, this embodiment discloses an online price assessment system based on big data-driven ship depreciation rate curve fitting, used to accurately value ship assets of different types, ages, construction locations, and navigation conditions. The system comprehensively utilizes historical transaction data, broker quotes, and multi-source market indicators, combined with expert fuzzy algorithms and deep regression models, to perform hierarchical modeling of the valuation process. By establishing a multi-layered mapping relationship from actual transaction prices to final assessed prices, it achieves adaptive response to market fluctuations and cross-cycle price correction.

[0054] The overall system structure consists of five hierarchical modules: 1. Data Acquisition Module (101): Used to collect real transaction data, broker quotes, basic ship information and macro market indicators from a global ship database; 2. Comprehensive Price Reverse Calculation Module (102): Used to perform multiple corrections on the transaction price and reverse calculate the comprehensive price under standard conditions; 3. Benchmark point generation module (103): Calculates the benchmark base by aggregating the comprehensive prices of ships of the same type and age; 4. Benchmark value fitting module (104): Fits a continuous benchmark value curve using a big data regression model; 5. Evaluation Calculation Module (105): Calculates the final evaluation price based on the characteristic parameters and correction coefficients of the target vessel.

[0055] The system employs a fully automated numerical learning and correction mechanism, dynamically updating the data using a large-scale transaction sample. Expert fuzzy algorithms are used to normalize and estimate correction coefficients that cannot be directly quantified (such as classification society reputation, differences in shipbuilding technology at the builder's location, and operational complexity in the navigation area), thereby ensuring the stability and interpretability of the assessment results in complex market environments.

[0056] This embodiment discloses an online price assessment method based on big data-driven ship depreciation rate curve fitting. Its core lies in integrating multi-source data through a hierarchical modeling architecture and utilizing expert fuzzy algorithms and regression models to achieve dynamic and accurate ship value assessment. The implementation process mainly includes data preprocessing, expert scoring, fuzzy inference, price inverse calculation, parameter fitting, and audit calibration. Specifically, firstly, the raw data is cleaned and standardized, outliers are removed, and units are unified; then, unstructured features are initially scored using an expert system, and correction coefficients are calculated based on fuzzy membership functions; next, the standardized comprehensive price is derived using the inverse calculation module, and a continuous value curve is constructed through the benchmark point generation and fitting module; finally, the valuation result of the target ship is output through the assessment module, and the model parameters are dynamically optimized based on a closed-loop update mechanism. The entire process is data-driven at its core, combined with artificial intelligence technology, to ensure the accuracy, real-time performance, and interpretability of the assessment results. The implementation process is as follows: (1) Sample preprocessing: Remove extreme outliers and... , Standardize the units.

[0057] (2) Initial expert scoring: The expert group conducted preliminary scoring mapping for the classification society, place of construction, and navigation area. .

[0058] (3) Membership calculation: Calculate the membership vector for each sample.

[0059] (4) Fuzzy inference aggregation: Calculated using fuzzy algorithms. .

[0060] (5) Comprehensive price inverse calculation: Calculation .

[0061] Expert System Parameters: The expert fuzzy system constructed in this invention uses basic ship attributes, life cycle characteristics, construction background, and operating environment as core input factors. Through systematic interval division of multi-dimensional parameters, it achieves valuation correction and state identification for different types of ships. The system's parameter system consists of five dimensions: ship type, ship age, classification society, place of construction, and navigation area. Each dimension is independently divided based on industry standards, market practices, and ship life cycle patterns, and corresponding membership functions and rule sets are configured in the fuzzy system to ensure sufficient coverage and differentiation capabilities when processing different ship samples.

[0062] In terms of ship type, this system further subdivides into three major categories: international bulk carriers, international oil tankers, and international container ships. It then groups these ships continuously based on tonnage or TEU ranges, providing a precise classification basis for major ship types and specific market segments. In terms of ship age, based on ship depreciation patterns and asset return characteristics at different life stages, the system divides ship age into seven intervals, including typical life segments such as newbuilds, in-service ships, and ships nearing the end of their service life, ensuring that the fuzzy system can continuously map the impact of ship age.

[0063] At the classification society level, this system covers major domestic and international classification organizations and incorporates flags of convenience and other categories to fully reflect the market differences arising from varying inspection standards. At the place of construction level, this system employs a "dual-dimensional place of construction coefficient system," setting independent correction coefficients for both the year of construction and ship type dimensions. This comprehensively covers China, Japan, South Korea, the Philippines, and other regions, simultaneously depicting price fluctuations throughout the shipbuilding cycle and differences in ship type development processes. At the navigation area level, this system distinguishes between international, coastal, and inland waterway navigation areas to reflect differences in structural strength requirements, equipment standards, and market prices corresponding to different operating environments.

[0064] Through the parameter system of the above five dimensions, this invention realizes a structured expression of the entire process of ship construction, service and market transaction, enabling the expert fuzzy system to perform consistent, scalable and interpretable valuation and classification of ships of different types, backgrounds and operating conditions, laying a complete data foundation for subsequent rule base construction, membership degree calculation and final valuation output.

[0065] 1. Ship type grouping: The ship types used in the calculations are mainly divided into three categories: international bulk carriers, international oil tankers, and international container ships, as detailed below: International bulk carriers are shown in Table 2.

[0066] Table 2 International oil tankers are shown in Table 3.

[0067] Table 3 International container ships are shown in Table 4.

[0068] Table 4 2. Ship age grouping: The calculation takes into account the different price curves of ships of different ages, and is grouped by age, specifically into 0-year ships, 5-year ships, 10-year ships, 15-year ships, 20-year ships, 25-year ships, and ships nearing scrap, as shown in Table 5.

[0069] Table 5 3. Classification Society Grouping: The classification society system covers major domestic and international organizations, encompassing more than ten categories including ZC, CCS (domestic and international), ABS, BV, NK, KR, RINA, LR, DNV, IRS, RS, and flags of convenience. Each classification society has its own independent correction factors, as shown in Table 6.

[0070] Table 6 4. Grouping by construction location (by construction date): To accurately reflect the impact of different construction locations on the price of newly built ships, this model divides the main construction regions into China, Japan, South Korea, the Philippines, and other regions. Furthermore, to reflect the interactive influence of the "construction year" and "ship age" dimensions on the construction location coefficient, the following correction coefficient rules are set: For each specific construction year (e.g., 2025, 2024) and each specific ship age (e.g., 0-year-old ship, 1-year-old ship), a complete set of construction location correction coefficients is independently configured. This set of coefficients comprehensively covers all the aforementioned regions, as shown in Table 7.

[0071] Table 7 Construction location grouping (by ship type): Building upon the previously defined "year of construction - age" dimension coefficient, this model adds an independent ship type dimension coefficient for the construction location. This coefficient also fully covers China, Japan, South Korea, the Philippines, and other regions. It aims to accurately quantify the unique price impact of different ship types (such as Panamax and Handysize) in different construction locations due to differences in design, technology, and market supply and demand, as shown in Table 8.

[0072] Table 8 5. Flight Area Grouping: To address the differences between vessels operating on international routes and those operating on domestic coastal or inland waterway routes, a navigation area correction factor is added, as shown in Table 9.

[0073] Table 9 To illustrate the specific implementation process of this invention in detail, the following uses a target vessel as an example to demonstrate the calculation process and output results of each module step by step. The basic parameters of the target vessel are as follows: Vessel type: New Panamax bulk carrier (tonnage range: 80,000-86,000 deadweight tons), deadweight tons: 82,000 tons, year of construction: 2018, current year: 2025 (age: 7 years), classification society: ABS, place of construction: Japan, navigation area: international.

[0074] Data Acquisition Module: Input data includes historical transaction records, broker quotes, and basic vessel information. Assume the following relevant data is collected: Actual transaction price ( ): A set of historical transaction samples similar to those of New Panamax bulk carriers, for example: Sample 1: =32 million US dollars (Transaction date: January 2024); Sample 2: =31 million US dollars (Transaction date: June 2024); Broker's quote ( ): The broker's quote corresponding to the sample, for example: Sample 1: =33 million US dollars, Sample 2: =31.5 million US dollars; Depreciation rate function parameters: based on historical data fitting, assuming... =0.05 (depreciation rate for the 0-10 year period). =0.08 (depreciation rate over 10 years). =10 years (accelerated depreciation point).

[0075] Initial expert assessment: The expert panel provides preliminary assessments based on the classification society, place of construction, and navigation area. ∈[0,1]), for example: ABS classification society: =0.9 (High Reputation), Built in Japan: =0.85 (high-quality craftsmanship), International Navigation Area: =0.8 (standard operation).

[0076] Comprehensive Price Inverse Calculation Module: The goal of this module is to calculate the standardized composite price. First, calculate the standardized transaction ratio r for each sample: r = 3200 / 3300 ≈ 0.9697; Next, multiple adjustments are made using the depreciation rate function and correction factors. The depreciation rate function f_dep(Y) adopts a piecewise exponential form: ; For ship age (10 years) ; For ship age , ; For this sample, ; The correction coefficient was then calculated using fuzzy inference. , , The membership function adopts a trapezoidal form, and the parameters are set by the expert group. The classification society correction factor is used. For example, let's set the membership function parameters as follows: =0.6, =0.7, =0.9, =1.0 (corresponding to the "high reputation" range); for =0.9, the membership degree is calculated as follows: Take the upper limit of 1.0; According to the formula Calculation by integration : Integrating over the interval [0,1] yields the following result: ≈0.92.

[0077] Similarly, calculate the construction site correction factor. ( =0.85) and flight area correction factor ( =0.8): ≈0.88 ≈0.82.

[0078] Now, calculate the value for each sample. : Sample 1: =3200 / (82000 0.7408 0.92 0.88 0.82)≈3200 / 39989.5≈0.0800 (unit: US$10,000 / ton); Sample 2: =3100 / (82000 0.6703 0.92 0.88 0.82)≈3100 / 36201.2≈0.0856.

[0079] Reference point generation: This module aggregates ships of the same type and age range. Value, generating the base number of the reference point Assume there are multiple samples of New Panamax bulk carriers within the 5-10 year age range, including Sample 1 and Sample 2 mentioned above. Sample Set ={ , ,...}, weight Based on time decay function ; in, This indicates the interval (in years) between the sample and the current time (2025). This is the time decay coefficient.

[0080] Sample 1: =1.5 years, =exp(-0.1 1.5)≈0.8607; Sample 2: =0.5 years, =exp(-0.1 0.5)≈0.9512; Baseline Calculated as a weighted average: =(0.0800 0.8607 + 0.0856 0.9512) / (0.8607+0.9512)≈(0.0689+0.0814) / 1.8119≈0.0829; Benchmark value fitting: This module is based on discrete reference points. Fitting a continuous benchmark function Assuming that for New Panamax bulk carriers, the benchmark points for different age groups are as follows: 5 years old: B=0.0900, 10 years old: B=0.0829, 15 years old: B=0.0700.

[0081] A spline regression model was employed. The parameters were optimized using gradient descent by minimizing the objective function. The fitted function was then analyzed at 7 years of ship age. ≈0.0850.

[0082] Assessment Calculation Module: This module calculates the final assessed value of the target vessel. Enter the target vessel parameters: deadweight tons: =82,000 tons; Age: =7 years; Classification society correction factor: =0.92 (updated in real time via fuzzy inference); Construction site correction coefficient: =0.88; Flight area correction factor: =0.82; Benchmark unit value: ≈0.0850; = 46,279,900 (USD); therefore, the appraised value of the target vessel is approximately USD 46,280,000.

[0083] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. An online price evaluation method based on big data ship depreciation rate curve fitting, characterized in that, include: Step 1: Obtain multi-dimensional ship information, and use the multi-dimensional ship information to make multiple corrections to the transaction price to obtain a comprehensive price under standard conditions, i.e., the standard price; Step 2: Aggregate the standard prices corresponding to the same ship type and age range to calculate the benchmark base, and fit a continuous benchmark curve by using a big data regression model with the goal of minimizing the fitting objective function. Step 3: Based on the benchmark curve, conduct a comprehensive assessment of the market value of the target vessel to obtain the final assessed price.

2. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, The multi-dimensional ship information includes: ship transaction price data, broker quote data, basic ship information, and macroeconomic market indicators.

3. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, Obtaining the standard price includes: ； in, For standard price, This is the actual transaction price. For deadweight tons, It is a function of depreciation rate. For classification society correction factors, For the construction site correction factor, This is the correction factor for the flight area.

4. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 3, characterized in that, The depreciation rate function includes: ； in, , For depreciation rate parameters, Year of construction The constants defined for the fuzzy expert method.

5. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, The benchmark base for calculating the standard price corresponding to the same ship type and age range includes: in, For the same ship type Same ship age The base number of the reference point, For standard price, As weight, For the sample set, , representing each sample.

6. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, The objective function for fitting is defined as follows: ； in, To fit the objective function, As the base point, For smoothing parameters, As a baseline function, for, This is the differential symbol.

7. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, Fitting a continuous baseline curve using the big data regression model includes: ； in, For B-spline basis functions, For learnable parameters, The total number of B-spline basis functions used. Number the basis functions.

8. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, Obtaining the final appraised price includes: ； in, For the target vessel's deadweight tonnage; The age of the target vessel; These are the latest correction coefficients for the target ship obtained through expert fuzzy reasoning; This is the base unit value for the corresponding ship age.

9. The online price evaluation method based on big data ship depreciation rate curve fitting according to claim 1, characterized in that, The method also includes: A closed-loop update mechanism is used to update multi-dimensional ship information in real time, and steps 1-2 are repeated until a new baseline curve is obtained.

10. An online price evaluation system based on big data ship depreciation rate curve fitting, implemented according to any one of claims 1-9, characterized in that, include: The data acquisition module is used to obtain multi-dimensional ship information; The comprehensive price inverse calculation module is used to perform multiple corrections on the transaction price using the multi-dimensional ship information to obtain the comprehensive price under standard conditions, i.e., the standard price. The benchmark generation module is used to aggregate the benchmark base for calculating standard prices corresponding to the same ship type and age range. The benchmark fitting module is used to fit a continuous benchmark curve using a big data regression model with the goal of minimizing the fitting objective function. The assessment calculation module is used to comprehensively assess the market value of the target vessel based on the benchmark curve and obtain the final assessment price.

Images