Price prediction method and device, computing equipment and computer readable medium

Abstract

The invention discloses a price prediction method and apparatus, a computing device and a computer readable storage medium. The method comprises the steps of obtaining time sequence data related to the price of a target material and initial material data; performing multi-modal feature fusion processing on the time sequence data and the initial material data to obtain multi-modal fusion features; inputting the multi-modal fusion features into a trained price prediction model to obtain a price prediction result of the target material output by the price prediction model; the price prediction model is established based on historical time sequence data, historical initial material data and corresponding historical prices. Thus, the time sequence data related to the price of the target material and the initial material data are subjected to multi-modal feature fusion processing to predict the price of the target material, the prediction accuracy and rapidness of the material price can be improved, and meanwhile the cost control ability and market competitiveness of an enterprise can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering machinery, and in particular to a price prediction method, device, computing equipment and computer-readable storage medium. Background Art

[0002] In the field of engineering machinery, the procurement cost of materials accounts for more than 85% of the total cost of products. The procurement of materials usually involves steel, alloys, hydraulic parts, etc. The price fluctuations of these materials have a significant impact on production costs and market pricing. Traditional material price forecasting methods rely on historical data and manual experience, which are difficult to cope with the frequent fluctuations in market prices, resulting in delayed data updates and inability to reflect market price fluctuations in real time, making forecasts inaccurate. At the same time, due to reasons such as data dispersion and low concentration, price forecasting takes a long time, resulting in delayed pricing, which affects the company's cost control capabilities and market competitiveness. Summary of the invention

[0003] The purpose of this application is to provide a price forecasting method, device, computing equipment and computer-readable storage medium, which can improve the accuracy and speed of material price forecasting, and at the same time improve the cost control ability and market competitiveness of the enterprise.

[0004] To achieve the above objectives:

[0005] In a first aspect, an embodiment of the present application provides a price prediction method, comprising:

[0006] Obtaining time series data related to the price of target materials and initial material data;

[0007] Performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features;

[0008] The multimodal fusion features are input into a trained price prediction model to obtain a price prediction result for the target material output by the price prediction model; the price prediction model is established based on historical time series data, historical initial material data and corresponding historical prices.

[0009] Optionally, the time series data includes at least one of the following: historical price data, macroeconomic data, industry supply and demand data, and financial market data; the initial material data includes at least one of the following: market analysis reports, corporate financial statements, and policy and regulatory documents.

[0010] Optionally, performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features includes:

[0011] Performing data enhancement processing on the time series data to obtain time series enhanced data;

[0012] Multimodal feature fusion is performed on the time series enhanced data and the initial material data to obtain multimodal fusion features.

[0013] Optionally, the performing multimodal feature fusion on the time series enhanced data and the initial material data to obtain multimodal fusion features includes:

[0014] Using a time-domain convolutional network algorithm to extract time-domain features from the time series enhanced data to obtain time series features;

[0015] Use the BERT model to perform deep semantic feature extraction on the initial material data to obtain initial material features;

[0016] Multimodal feature fusion is performed on the time series features and the initial material features to obtain multimodal fusion features.

[0017] Optionally, performing data enhancement processing on the time series data to obtain time series enhanced data includes:

[0018] Decomposing the time series data by using a time series decomposition algorithm to obtain a trend component for representing the long-term change trend of the time series, a seasonal component for representing the periodic fluctuation or seasonal change of the time series, and a residual component for random fluctuations that cannot be explained by the trend component or the seasonal component;

[0019] The trend component, the seasonal component and the residual component are reconstructed to generate time series enhanced data.

[0020] Optionally, reconstructing the trend component, the seasonal component and the residual component to generate time series enhanced data includes:

[0021] Reconstructing the trend component based on the historical trend component to generate a new trend component;

[0022] Generate a new seasonal component according to the historical seasonal component or by adding random disturbance to the seasonal component;

[0023] Performing random sampling or noise simulation on the residual component to generate a new residual component;

[0024] The generated new trend component, the new seasonal component and the new residual component are added together to obtain time series enhanced data.

[0025] Optionally, the price prediction result includes a predicted price of the target material; and the method further includes:

[0026] According to the BOM management list of the target product, determine the process and required quantity required to process the target material into at least one target component; the target component is a component in the BOM management list that is related to the target material;

[0027] Determine the labor costs required to process the target material into each of the target components according to the process;

[0028] The cost price of the target product with respect to the target material is determined based on the predicted price of the target material, the labor cost, and the demand quantity.

[0029] In a second aspect, an embodiment of the present application provides a price prediction device, comprising:

[0030] An acquisition module, for acquiring time series data related to the price of a target material and initial material data;

[0031] A processing module, used for performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features;

[0032] A prediction module is used to input the multimodal fusion features into a trained price prediction model to obtain a price prediction result of the target material output by the price prediction model; the price prediction model is established based on historical time series data, historical initial material data and corresponding historical prices.

[0033] In a third aspect, an embodiment of the present application provides a computing device, comprising: a processor and a memory storing a computer program, and when the processor runs the computer program, the above-mentioned price prediction method is implemented.

[0034] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the above-mentioned price prediction method is implemented.

[0035] The price prediction method, device, computing device and computer-readable storage medium provided by the embodiment of the present application include: obtaining time series data and initial material data related to the price of the target material; performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features; inputting the multimodal fusion features into the trained price prediction model to obtain the price prediction result of the target material output by the price prediction model; the price prediction model is established based on the historical time series data and the historical initial material data and the corresponding historical prices. In this way, by combining the time series data related to the price of the target material and the initial material data related to the price of the target material for price prediction, a richer and higher-quality data set is provided for material price prediction, and by performing multimodal feature fusion on the time series data and the initial material data, the potential relationship between the time series data and the initial material data is captured as much as possible and multi-source feature information is aggregated, so as to improve the accuracy, comprehensiveness and speed of material price prediction, and at the same time improve the cost control ability and market competitiveness of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of a flow chart of a price prediction method provided by an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of the architecture of a price prediction method provided by an embodiment of the present invention;

[0038] Figure 3 Schematic diagram of a time series decomposition algorithm according to an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of the structure of a price prediction device provided by an embodiment of the present invention;

[0040] Figure 5 A schematic diagram of the structure of a computing device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0042] It should be noted that, in this article, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, components, features, and elements with the same name in different embodiments of the present application may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context of the specific embodiment.

[0043] It should be understood that, although the terms first, second, third, etc. may be used to describe various information in this article, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this article, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein can be interpreted as "at the time of..." or "when..." or "in response to determination". Furthermore, as used in this article, the singular forms "one", "one" and "the" are intended to also include plural forms, unless there is an opposite indication in the context. It should be further understood that the terms "comprising", "including" indicate that there are described features, steps, operations, elements, components, projects, kinds, and / or groups, but do not exclude the existence, occurrence or addition of one or more other features, steps, operations, elements, components, projects, kinds, and / or groups. The terms "or" and "and / or" used herein are interpreted as inclusive, or mean any one or any combination. Thus, “A, B, or C” or “A, B and / or C” means “any of the following: A; B; C; A and B; A and C; B and C; A, B, and C.” An exception to this definition will occur only when a combination of elements, functions, steps, or operations are inherently mutually exclusive in some manner.

[0044] It should be understood that, although the various steps in the flowchart in the embodiment of the present application are displayed in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and it can be performed in other orders. Moreover, at least a portion of the steps in the figure may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and their execution order is not necessarily performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0045] It should be noted that, in this article, step codes such as S101, S102, etc. are used for the purpose of expressing the corresponding content more clearly and concisely, and do not constitute a substantial limitation on the sequence. When implementing the step, those skilled in the art may execute S102 first and then S101, etc., but these should all be within the scope of protection of this application.

[0046] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0047] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present application, and have no specific meanings. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0048] See also Figure 1 and Figure 2 The embodiment of the present application provides a price prediction method, which can be executed by a price prediction device provided by the embodiment of the present application. The price prediction device can be implemented in software and / or hardware, such as a computing device such as a computer or a server. In this embodiment, the execution subject of the price prediction method is a computing device as an example. The price prediction method provided by this embodiment includes:

[0049] Step S101: Acquire time series data related to the price of a target material and initial material data.

[0050] Among them, the target material is the material that needs to be price predicted, which can be set based on actual needs. It should be noted that the material in this embodiment can refer to raw materials, such as aluminum ore, etc., or it can refer to materials obtained after processing raw materials, such as aluminum foil, aluminum alloy, etc. Among them, affected by factors such as the industrial chain and the environment, there can be multiple time series data and initial material data related to the price of the target material. In one embodiment, the time series data includes at least one of the following: historical price data, macroeconomic data, industry supply and demand data, and financial market data; the initial material data includes at least one of the following: market analysis reports, corporate announcement financial reports, and policy and regulatory documents. Historical price data can be price data within a period of time before the current time, such as price data for the last six months or three months. Macroeconomic data can refer to data such as GDP and price index of a country or region. Industry supply and demand data can refer to data such as demand and supply of target materials in the industry where the target material is located. Financial market data can refer to price change data of target materials in the financial market or a specific area. It is understandable that changes in macroeconomic data, industry supply and demand data, and financial market data may affect the price fluctuations of target materials, which may lead to higher or lower prices of target materials in different quarters or at different times. Among them, the market analysis report may refer to an analysis report on the market demand, market price, etc. of the target materials. The company's public financial report may be the company's public financial report in the past period of time, such as half a year or a year, which can reflect the sales trend of the products of the target materials, and based on the sales trend of the product, the subsequent demand trend for the target materials can be predicted, such as an increase or decrease in demand. Policy and regulatory documents may refer to policies and regulations on the production, processing, etc. of the target materials. It is understandable that when the market analysis report, the company's public financial report and / or the policy and regulatory documents change, it may affect the price of the target materials.

[0051] Step S102: Perform multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features.

[0052] Specifically, data enhancement processing is performed on the time series data to obtain time series enhanced data; multimodal feature fusion is performed on the time series enhanced data and the initial material data to obtain multimodal fusion features.

[0053] It can be understood that by fusing the time series data and the initial material data with multimodal features, a more comprehensive perspective can be provided for material price forecasting, and the advantages of different data sources can be combined to capture the potential relationship between the time series data and the initial material data as much as possible and aggregate multi-source feature information, so that the obtained multimodal fusion features contain information that plays an important role in subsequent forecasts.

[0054] In one embodiment, performing data enhancement processing on time series data to obtain time series enhanced data includes:

[0055] Decompose the time series data using the time series decomposition algorithm to obtain the trend component used to represent the long-term trend of the time series, the seasonal component used to represent the periodic fluctuation or seasonal change of the time series, and the residual component used for the random fluctuation that cannot be explained by the trend component or the seasonal component;

[0056] The trend component, seasonal component and residual component are reconstructed to generate time series enhanced data.

[0057] Among them, Figure 3 As shown, the sliding window and local regression techniques in the time series decomposition (STL) algorithm can be used to decompose the time series data to decompose the time series data into trend components (also called trend components), seasonal components (also called seasonal components) and residual components (also called residual components). Then, the trend components, seasonal components and residual components are reconstructed to generate time series enhanced data with more actual market volatility, thereby improving the accuracy, comprehensiveness and speed of subsequent material price forecasts.

[0058] In one embodiment, reconstructing the trend component, the seasonal component, and the residual component to generate time series enhanced data includes:

[0059] Reconstruct the trend component based on the historical trend component to generate a new trend component;

[0060] Generate new seasonal components based on historical seasonal components or by adding random disturbances to seasonal components;

[0061] Perform random sampling or noise simulation on the residual components to generate new residual components;

[0062] The generated new trend component, new seasonal component and new residual component are added together to obtain time series enhanced data.

[0063] Among them, based on the historical trend component, the trend component can be reconstructed by applying a regression model or smoothing technology to generate a new trend component. At the same time, a new seasonal component is generated based on the existing seasonal component, that is, the historical seasonal component, for example, a new seasonal component is generated based on the difference between the historical seasonal component and the current seasonal component, or a new seasonal component can be generated by adding random perturbations to the seasonal component. In addition, a new residual component is generated by randomly sampling or noise simulation of the residual component to reflect the random volatility of prices in the market.

[0064] In one embodiment, multimodal feature fusion is performed on the time series enhanced data and the initial material data to obtain multimodal fusion features, including:

[0065] Use the time domain convolutional network algorithm to extract time domain features from time series enhanced data to obtain time series features;

[0066] Use the BERT model to extract deep semantic features from the initial material data to obtain initial material features;

[0067] Multimodal feature fusion is performed on time series features and initial material features to obtain multimodal fusion features.

[0068] Among them, the time-domain feature extraction of time series enhanced data through the time-domain convolutional network algorithm can capture the time series features that can be used to characterize price trends and price fluctuations, and the deep semantic feature extraction of initial material data through the BERT model can obtain initial material features that can reflect market sentiment and policy changes. Then, by multimodal feature fusion of time series features and initial material features, we can focus on the most important information for prediction and obtain multimodal fusion features. In this way, the accuracy, comprehensiveness and speed of material price prediction are further improved.

[0069] Step S103: input the multimodal fusion features into the trained price prediction model to obtain the price prediction result of the target material output by the price prediction model; the price prediction model is established based on historical time series data, historical initial material data and corresponding historical prices.

[0070] Specifically, the obtained multimodal fusion features are input into the trained price prediction model, so that the price prediction model predicts the price of the target material based on the input multimodal fusion features, and obtains the price prediction result of the target material output by the price prediction model, which may include the predicted price of the target material in a period of time in the future.

[0071] Among them, the price prediction model can be built using the residual network (ResNet) algorithm, and trained based on historical time series data, historical initial material data and corresponding historical prices. The specific training process can refer to the existing technology and will not be repeated here. Here, the residual network can learn complex feature patterns by nonlinearly mapping the fused multimodal features, and then generate material price prediction results based on the fused multimodal features.

[0072] In summary, in the price prediction method provided in the above embodiment, price prediction is performed by combining time series data related to the price of the target material and initial material data related to the price of the target material, so as to provide a richer and higher-quality data set for material price prediction. At the same time, multimodal feature fusion is performed on the time series data and the initial material data to capture the potential relationship between the time series data and the initial material data as much as possible and aggregate multi-source feature information, thereby improving the accuracy, comprehensiveness and speed of material price prediction, and at the same time improving the cost control ability and market competitiveness of the enterprise.

[0073] In one embodiment, the price prediction result includes a predicted price of the target material; and the method further includes:

[0074] According to the BOM management list of the target product, determine the process and quantity required to process the target material into at least one target component; the target component is the component related to the target material in the BOM management list;

[0075] Determine the cost of processing the target material into each target component according to the process;

[0076] Based on the predicted price, labor cost and demand of the target material, the cost price of the target product regarding the target material is determined.

[0077] Among them, the target product can be engineering machinery and equipment, etc., which is not specifically limited here. The BOM management list of the target product may include all the parts information required for the production or manufacture of the target product, and the parts information may include the quantity of each part, the materials constituting each part, and the process and demand for processing each material into the corresponding part. Therefore, according to the BOM management list of the target product, the target parts related to the target material and the process required to process the target material into the corresponding target part can be determined first. Since the price of each process is relatively fixed, that is, equivalent to known, the labor cost required for processing the target material into each target part can be determined according to the process. In addition, under the conditions of the known predicted price of the target material, the labor cost required for processing the target material into each target part, and the demand for processing the target material into the corresponding part, the cost price of the target product for the target material can be calculated. Here, the process compilation is obtained through the BOM management list, and the labor cost price corresponding to the compliant product is generated by processing the target material, and then the predicted price and labor cost price of the target material are jointly composed of the cost price of the target product for the target material. In this way, by obtaining the cost price of the target product for the target material, the enterprise's responsiveness to market changes can be enhanced, and the product competitiveness can be improved.

[0078] Based on the same inventive concept as the above embodiments, Figure 4As shown, an embodiment of the present invention provides a price prediction device, comprising:

[0079] An acquisition module, for acquiring time series data related to the price of a target material and initial material data;

[0080] A processing module, used for performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features;

[0081] The prediction module is used to input the multimodal fusion features into the trained price prediction model to obtain the price prediction result of the target material output by the price prediction model; the price prediction model is established based on historical time series data, historical initial material data and corresponding historical prices.

[0082] In one embodiment, the time series data includes at least one of the following: historical price data, macroeconomic data, industry supply and demand data, and financial market data; the initial material data includes at least one of the following: market analysis reports, corporate financial statements, and policy and regulatory documents.

[0083] In one embodiment, the processing module is specifically configured to:

[0084] Performing data enhancement processing on the time series data to obtain time series enhanced data;

[0085] Multimodal feature fusion is performed on the time series enhanced data and the initial material data to obtain multimodal fusion features.

[0086] In one embodiment, the processing module is specifically configured to:

[0087] Using a time-domain convolutional network algorithm to extract time-domain features from the time series enhanced data to obtain time series features;

[0088] Use the BERT model to perform deep semantic feature extraction on the initial material data to obtain initial material features;

[0089] Multimodal feature fusion is performed on the time series features and the initial material features to obtain multimodal fusion features.

[0090] In one embodiment, the processing module is specifically configured to:

[0091] Decomposing the time series data by using a time series decomposition algorithm to obtain a trend component for representing the long-term change trend of the time series, a seasonal component for representing the periodic fluctuation or seasonal change of the time series, and a residual component for random fluctuations that cannot be explained by the trend component or the seasonal component;

[0092] The trend component, the seasonal component and the residual component are reconstructed to generate time series enhanced data.

[0093] In one embodiment, the processing module is specifically configured to:

[0094] Reconstructing the trend component based on the historical trend component to generate a new trend component;

[0095] Generate a new seasonal component according to the historical seasonal component or by adding random disturbance to the seasonal component;

[0096] Performing random sampling or noise simulation on the residual component to generate a new residual component;

[0097] The generated new trend component, the new seasonal component and the new residual component are added together to obtain time series enhanced data.

[0098] In one embodiment, the price prediction result includes a predicted price of the target material;

[0099] The acquisition module is further used to determine the process and required quantity required to process the target material into at least one target component according to the BOM management list of the target product; the target component is a component related to the target material in the BOM management list;

[0100] The processing module is further used to determine the labor costs required to process the target material into each of the target components according to the process;

[0101] The prediction module is also used to determine the cost price of the target product with respect to the target material based on the predicted price of the target material, the labor cost and the demand.

[0102] Based on the same inventive concept as the above embodiments, an embodiment of the present invention provides a computing device, such as Figure 5 As shown, the computing device includes: a processor 310 and a memory 311 storing a computer program; wherein, Figure 5 The processor 310 shown in the figure is not used to indicate that the number of the processor 310 is one, but is only used to indicate the positional relationship of the processor 310 relative to other devices. In actual applications, the number of the processor 310 may be one or more; similarly, Figure 5 The memory 311 shown in the figure has the same meaning, that is, it is only used to refer to the position relationship of the memory 311 relative to other devices. In practical applications, the number of memories 311 can be one or more. When the processor 310 runs the computer program, the price prediction method described above is implemented.

[0103] The computing device may also include: at least one network interface 312. The various components in the computing device are coupled together via a bus system 313. It is understood that the bus system 313 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 313 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the bus system 313 is not used in the example above. Figure 5 Various buses are labeled as bus system 313.

[0104] The memory 311 may be a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memories. The nonvolatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM); the magnetic surface memory may be a disk memory or a tape memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM, SyncLink Dynamic Random Access Memory), and direct RAM bus random access memory (DRRAM, Direct Rambus Random Access Memory).The memory 311 described in the embodiments of the present invention is intended to include but is not limited to these and any other suitable types of memories.

[0105] The memory 311 in the embodiment of the present invention is used to store various types of data to support the operation of the computing device. Examples of these data include: any computer program used to operate on the computing device, such as an operating system and an application; contact data; phone book data; messages; pictures; videos, etc. Among them, the operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application program may include various applications, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. Here, the program that implements the method of the embodiment of the present invention may be included in the application program.

[0106] Based on the same inventive concept as the above-mentioned embodiment, this embodiment further provides a computer storage medium, wherein a computer program is stored in the computer storage medium, and the computer storage medium may be a magnetic random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM) or other memory; or may be various devices including one or any combination of the above-mentioned memories, such as a mobile phone, a computer, a tablet device, a personal digital assistant, etc. When the computer program stored in the computer storage medium is executed by the processor, the price prediction method described above is implemented. For the specific steps implemented when the computer program is executed by the processor, please refer to Figure 1 The description of the illustrated embodiment will not be repeated here.

[0107] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0108] In this document, the terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of elements other than those listed and may also include additional elements not expressly listed.

[0109] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A price prediction method, characterized in that: The method comprises: Obtaining time series data related to the price of target materials and initial material data; Performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features; The multimodal fusion features are input into a trained price prediction model to obtain a price prediction result for the target material output by the price prediction model; the price prediction model is established based on historical time series data, historical initial material data and corresponding historical prices.

2. The method according to claim 1, characterized in that: The time series data includes at least one of the following: historical price data, macroeconomic data, industry supply and demand data, and financial market data; the initial material data includes at least one of the following: market analysis reports, corporate financial statements, and policy and regulatory documents.

3. The method according to claim 1 or 2, characterized in that: The performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features includes: Performing data enhancement processing on the time series data to obtain time series enhanced data; Multimodal feature fusion is performed on the time series enhanced data and the initial material data to obtain multimodal fusion features.

4. The method according to claim 3, characterized in that: The performing multimodal feature fusion on the time series enhanced data and the initial material data to obtain multimodal fusion features includes: Using a time-domain convolutional network algorithm to extract time-domain features from the time series enhanced data to obtain time series features; Use the BERT model to perform deep semantic feature extraction on the initial material data to obtain initial material features; Multimodal feature fusion is performed on the time series features and the initial material features to obtain multimodal fusion features.

5. The method according to claim 3, characterized in that: The performing data enhancement processing on the time series data to obtain time series enhanced data includes: Decomposing the time series data by using a time series decomposition algorithm to obtain a trend component for representing the long-term change trend of the time series, a seasonal component for representing the periodic fluctuation or seasonal change of the time series, and a residual component for random fluctuations that cannot be explained by the trend component or the seasonal component; The trend component, the seasonal component and the residual component are reconstructed to generate time series enhanced data.

6. The method according to claim 5, characterized in that The step of reconstructing the trend component, the seasonal component and the residual component to generate time series enhanced data includes: Reconstructing the trend component based on the historical trend component to generate a new trend component; Generate a new seasonal component according to the historical seasonal component or by adding random disturbance to the seasonal component; Performing random sampling or noise simulation on the residual component to generate a new residual component; The generated new trend component, the new seasonal component and the new residual component are added together to obtain time series enhanced data.

7. The method according to claim 1, characterized in that The price prediction result includes the predicted price of the target material; the method further includes: According to the BOM management list of the target product, determine the process and required quantity required to process the target material into at least one target component; the target component is a component in the BOM management list that is related to the target material; Determine the labor costs required to process the target material into each of the target components according to the process; The cost price of the target product with respect to the target material is determined based on the predicted price of the target material, the labor cost, and the demand quantity.

8. A price prediction device, characterized in that: include: An acquisition module, for acquiring time series data related to the price of a target material and initial material data; A processing module, used for performing multimodal feature fusion processing on the time series data and the initial material data to obtain multimodal fusion features; A prediction module is used to input the multimodal fusion features into a trained price prediction model to obtain a price prediction result of the target material output by the price prediction model; the price prediction model is established based on historical time series data, historical initial material data and corresponding historical prices.

9. A computing device, characterized in that include: A processor and a memory storing a computer program, wherein when the processor runs the computer program, the price prediction method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the price prediction method according to any one of claims 1 to 7 is implemented.

Images