Refrigerating capacity prediction method and device based on transfer learning

Abstract

The invention relates to the technical field of refrigerating capacity prediction, and provides a refrigerating capacity prediction method and device based on transfer learning. The method comprises the following steps: acquiring first historical data of a first place and second historical data of a second place; performing coding and decoding training on the automatic coding and decoding network by using the first historical data, and performing refrigerating capacity prediction training on the prediction network; respectively freezing the weight of the last layer of full-connection network in the automatic coding and decoding network and the prediction network; performing coding and decoding training on the automatic coding and decoding network by using the second historical data, and performing refrigerating capacity prediction training on the prediction network; respectively unfreezing the weight of the last layer of full-connection network in the automatic coding and decoding network and the prediction network; and the automatic coding and decoding network and the prediction network are utilized to predict the refrigerating capacity required by the second place at different time of each day. By means of the technical means, the problems that in the prior art, due to the fact that the training data amount is small, refrigerating capacity prediction is over-fitted, and the generalization ability is poor are solved.

Description

Technical Field

[0001] The present application relates to the technical field of refrigeration capacity prediction, and in particular to a refrigeration capacity prediction method and device based on transfer learning. Background Art

[0002] Current cooling capacity prediction methods are mainly based on machine learning and statistical models, which usually require a large amount of historical data to train the models, such as support vector machines, decision trees, neural networks, etc. The core of these models is to predict future cooling demand by learning patterns in historical data. For example, in a shopping mall with sufficient data, data based on characteristics such as ambient temperature, humidity, and traffic flow can accurately predict cooling demand. However, when the amount of training data is small, the model is prone to overfitting and has poor generalization ability. Summary of the invention

[0003] In view of this, the embodiments of the present application provide a refrigeration capacity prediction method, device, electronic device and computer-readable storage medium based on transfer learning to solve the problem of overfitting and poor generalization ability of refrigeration capacity prediction caused by small amount of training data in the prior art.

[0004] According to a first aspect of an embodiment of the present application, a method for predicting cooling capacity based on transfer learning is provided, comprising: obtaining first historical data of a first location and second historical data of a second location, wherein the first historical data and the second historical data both comprise location data, environmental data, weather data and cooling capacity; performing encoding and decoding training on an automatic encoding and decoding network using the first historical data, and performing cooling capacity prediction training on a prediction network; freezing the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network, respectively; performing encoding and decoding training on the automatic encoding and decoding network and performing cooling capacity prediction training on the prediction network, respectively; unfreezing the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network, respectively; and predicting the cooling capacity required for the second location at different times of the day using the automatic encoding and decoding network and the prediction network.

[0005] According to a second aspect of an embodiment of the present application, a refrigeration capacity prediction device based on transfer learning is provided, comprising: an acquisition module, configured to acquire first historical data of a first place and second historical data of a second place, wherein the first historical data and the second historical data both comprise: place data, environmental data, weather data and refrigeration capacity; a first training module, configured to perform encoding and decoding training on an automatic encoding and decoding network using the first historical data, and to perform refrigeration capacity prediction training on a prediction network; a freezing module, configured to freeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network, respectively; a second training module, configured to perform encoding and decoding training on the automatic encoding and decoding network and to perform refrigeration capacity prediction training on the prediction network; a thawing module, configured to thaw the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network, respectively; and a prediction module, configured to predict the refrigeration capacity required for the second place at different times of the day using the automatic encoding and decoding network and the prediction network.

[0006] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0007] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0008] Compared with the prior art, the embodiments of the present application have the following beneficial effects: obtaining the first historical data of the first place and the second historical data of the second place, wherein the first historical data and the second historical data both include: place data, environmental data, weather data and cooling capacity; using the first historical data to perform encoding and decoding training on the automatic encoding and decoding network, and performing cooling capacity prediction training on the prediction network; freezing the weights of the last layer of fully connected networks in the automatic encoding and decoding network and the prediction network respectively; using the second historical data to perform encoding and decoding training on the automatic encoding and decoding network, and performing cooling capacity prediction training on the prediction network; unfreezing the weights of the last layer of fully connected networks in the automatic encoding and decoding network and the prediction network respectively; using the automatic encoding and decoding network and the prediction network to predict the cooling capacity required for the second place at different times of the day. The above-mentioned technical means can solve the problem in the prior art that the small amount of training data leads to overfitting of cooling capacity prediction and poor generalization ability, thereby avoiding overfitting of cooling capacity prediction and improving generalization ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0010] Figure 1 It is a flow chart of a refrigeration capacity prediction method based on transfer learning provided in an embodiment of the present application; Figure 2 It is a flow chart of a refrigeration capacity prediction training method provided in an embodiment of the present application; Figure 3 It is a structural schematic diagram of a refrigeration capacity prediction device based on transfer learning provided in an embodiment of the present application; Figure 4 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0011] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0012] A method and device for predicting refrigeration capacity based on transfer learning according to an embodiment of the present application will be described in detail below with reference to the accompanying drawings.

[0013] Figure 1 It is a flow chart of a refrigeration capacity prediction method based on transfer learning provided in an embodiment of the present application. Figure 1 The refrigeration capacity prediction method based on transfer learning can be executed by a computer or a server, or software on the computer or the server. Figure 1 As shown, the refrigeration capacity prediction method based on transfer learning includes: S101, obtaining first historical data of a first location and second historical data of a second location, wherein the first historical data and the second historical data both include: location data, environment data, weather data and cooling capacity; S102, performing encoding and decoding training on the automatic encoding and decoding network using the first historical data, and performing cooling capacity prediction training on the prediction network; S103, freezing the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; S104, performing encoding and decoding training on the automatic encoding and decoding network using the second historical data, and performing cooling capacity prediction training on the prediction network; S105, unfreeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; S106, using the automatic encoding and decoding network and the prediction network to predict the cooling capacity required by the second place at different times of the day.

[0014] The first place has multiple first historical data, and the second place has multiple second historical data, and the number of first historical data far exceeds the number of second historical data. The first place has abundant data, and the second place has less data. If the data of the second place is used to train the model applied to the second place, it will lead to problems of overfitting and poor generalization ability. In order to solve this problem, the embodiment of the present application first uses the first historical data of the first place to train the automatic encoding and decoding network and the prediction network, and then according to the principle of transfer learning, uses the second historical data of the second place to fine-tune the automatic encoding and decoding network and the prediction network. The automatic encoding and decoding network and the prediction network obtained in this way can be well applied in the second place.

[0015] The venue data includes the latitude and longitude of the venue, the floor area, the total building area of ​​the venue, the total number of shops in the venue, the distribution and proportion of the venue business types (the distribution and proportion of different types of shops). The environmental data includes the temperature, humidity, and flow of people inside the venue. The weather data includes the temperature, humidity, and wind speed outside the venue. The cooling capacity is the actual cooling capacity used in the venue data, environmental data, and weather data. It is used as a label in subsequent training.

[0016] According to the technical solution provided in the embodiment of the present application, the first historical data of the first place and the second historical data of the second place are obtained, wherein the first historical data and the second historical data both include: place data, environmental data, weather data and cooling capacity; the first historical data is used to perform encoding and decoding training on the automatic encoding and decoding network, and the prediction network is trained for cooling capacity prediction; the weights of the last layer of fully connected networks in the automatic encoding and decoding network and the prediction network are frozen respectively; the second historical data is used to perform encoding and decoding training on the automatic encoding and decoding network, and the prediction network is trained for cooling capacity prediction; the weights of the last layer of fully connected networks in the automatic encoding and decoding network and the prediction network are unfrozen respectively; the automatic encoding and decoding network and the prediction network are used to predict the cooling capacity required for the second place at different times of the day. The above-mentioned technical means can solve the problem in the prior art that the small amount of training data leads to overfitting of cooling capacity prediction and poor generalization ability, thereby avoiding overfitting of cooling capacity prediction and improving generalization ability.

[0017] In an optional embodiment, the historical data of the first shopping mall is collected, including the shopping mall's own dimension data, shopping mall environmental data, outdoor weather data, date data, and corresponding hourly cooling data. The discrete features are converted into category types and then converted into numerical values ​​for embedding, and the numerical features are normalized. The data sets are divided in chronological order, including training data sets, test data sets, and verification data sets. The network structure of the automatic encoding and decoding network is defined, the optimizer is determined, the learning rate, the number of iterations, and other model training are determined, and finally the automatic encoding and decoding network is trained for encoding and decoding using the first historical data, and the optimal model is selected for storage. The network structure of the prediction network is defined, the optimizer is determined, the learning rate, the number of iterations, and other model training are determined, and finally the prediction network is trained for cooling capacity prediction using the historical data of the first shopping mall, and the optimal model is selected for storage. The automatic encoding and decoding network and the prediction network are loaded, the weights of the last fully connected layer network of the automatic encoding and decoding network and the pre-trained model are frozen, the model is fine-tuned using the historical data of the second shopping mall, and the fine-tuned model is used to predict the cooling capacity of the second shopping mall data.

[0018] Furthermore, the first historical data is used to perform encoding and decoding training on the automatic encoding and decoding network, including: respectively processing the place data, environment data and weather data in the first historical data through the automatic encoding and decoding network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; respectively calculating the reconstruction losses between the place data and the place reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data using a mean square error loss function; and optimizing the weights in the automatic encoding and decoding network based on the reconstruction losses to complete the encoding and decoding training of the automatic encoding and decoding network using the first historical data.

[0019] Before respectively processing the location data, environment data and weather data in the first historical data through the automatic encoding and decoding network, the location data, environment data and weather data can be converted into embedded features, and the automatic encoding and decoding network reconstructs the embedded features.

[0020] During training, the plurality of first historical data may be divided into data sets in chronological order, including a training data set, a test data set and a verification data set. The automatic encoding and decoding network is trained using the training data set, and the effect of the automatic encoding and decoding network is tested and verified using the test data set and the verification data set.

[0021] Furthermore, the prediction network is trained for cooling capacity prediction using the first historical data, including: after the automatic encoding and decoding network is trained using the first historical data, the location data, environmental data and weather data in the first historical data are respectively processed by the automatic encoding and decoding network to obtain location reconstruction data, environmental reconstruction data and weather reconstruction data; the location reconstruction data, environmental reconstruction data and weather reconstruction data are processed by the prediction network to obtain a cooling capacity prediction value corresponding to the first historical data; the prediction loss between the cooling capacity prediction value and the cooling capacity in the first historical data is calculated using a regression loss function; and the weights in the prediction network are optimized based on the prediction loss to complete the cooling capacity prediction training of the prediction network using the first historical data.

[0022] The automatic encoding and decoding network reconstructs the embedded features, fuses the reconstructed features (place reconstruction data, environment reconstruction data and weather reconstruction data), processes the fusion results through the prediction network, and obtains the cooling capacity prediction value corresponding to the first historical data.

[0023] Furthermore, the second historical data is used to perform encoding and decoding training on the automatic encoding and decoding network, including: respectively processing the place data, environment data and weather data in the second historical data through the automatic encoding and decoding network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; respectively calculating the reconstruction losses between the place data and the place reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data using the mean square error loss function; and optimizing the weights in the automatic encoding and decoding network based on the reconstruction loss to complete the encoding and decoding training of the automatic encoding and decoding network using the second historical data.

[0024] Using the second historical data to perform codec training on the automatic codec network is similar to using the first historical data to perform codec training on the automatic codec network. The second historical data is used to fine-tune the automatic codec network based on the training of the automatic codec network with the first historical data.

[0025] Furthermore, the prediction network is trained for cooling capacity prediction using the second historical data, including: after the automatic encoding and decoding network is trained using the second historical data, the location data, environmental data and weather data in the second historical data are respectively processed by the automatic encoding and decoding network to obtain location reconstruction data, environmental reconstruction data and weather reconstruction data; the location reconstruction data, environmental reconstruction data and weather reconstruction data are processed by the prediction network to obtain a cooling capacity prediction value corresponding to the second historical data; the prediction loss between the cooling capacity prediction value and the cooling capacity in the second historical data is calculated using a regression loss function; and the weights in the prediction network are optimized based on the prediction loss to complete the cooling capacity prediction training of the prediction network using the second historical data.

[0026] Using the second historical data to train the prediction network for refrigeration capacity prediction is similar to using the first historical data to train the prediction network for refrigeration capacity prediction. The second historical data is used to fine-tune the prediction network based on the prediction network trained with the first historical data. The regression loss function may be Huber Loss.

[0027] Furthermore, the automatic encoding and decoding network and the prediction network are used to predict the cooling capacity required for the second place at different times of the day, including: obtaining the location data, environmental data, and weather data of the second place at the current moment; processing the location data, environmental data, and weather data of the second place at the current moment respectively through the automatic encoding and decoding network to obtain location reconstruction data, environmental reconstruction data, and weather reconstruction data; processing the location reconstruction data, environmental reconstruction data, and weather reconstruction data through the prediction network to obtain the cooling capacity prediction value of the second place at the current moment.

[0028] The embodiment of the present application is the reasoning stage of the automatic encoding and decoding network and the prediction network. Before the embodiment of the present application is the training stage of the automatic encoding and decoding network and the prediction network. The reasoning stage and the training stage of the network correspond to each other, and the reasoning stage will not be explained in detail here.

[0029] Furthermore, before performing encoding and decoding training on the automatic encoding and decoding network using the first historical data and performing refrigeration capacity prediction training on the prediction network, the method also includes: constructing the automatic encoding and decoding network using multiple layers of fully connected layers and activation functions; and constructing the prediction network using multiple layers of fully connected layers, batch normalization and activation functions.

[0030] The automatic encoding and decoding network is an AutoEncoder model, which includes an encoding part and a decoding part. The encoding part and the decoding part are constructed using multiple layers of fully connected layers and activation functions. The structures of the encoding part and the decoding part are the same, and they are used for encoding and decoding respectively. The prediction network is equivalent to a multi-layer perceptron (MLP).

[0031] Figure 2 is a flow chart of a cooling capacity prediction training method provided in an embodiment of the present application. Figure 2 As shown, the method includes: S201, respectively processing the place data, the environment data and the weather data in the first historical data through an automatic encoding and decoding network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; S202, using a mean square error loss function to respectively calculate the reconstruction losses between the location data and the location reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data; S203, optimizing the weights in the automatic encoding and decoding network according to the reconstruction loss, so as to complete the encoding and decoding training of the automatic encoding and decoding network by the first historical data; S204, after using the first historical data to perform codec training on the automatic codec network, respectively process the place data, environment data and weather data in the first historical data through the automatic codec network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; S205, obtaining a cooling capacity prediction value corresponding to the first historical data by processing the site reconstruction data, the environment reconstruction data and the weather reconstruction data through the prediction network; S206, calculating the prediction loss between the cooling capacity prediction value and the cooling capacity in the first historical data by using a regression loss function; S207, optimizing the weights in the prediction network according to the prediction loss, so as to complete the refrigeration capacity prediction training of the prediction network based on the first historical data.

[0032] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0033] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0034] Figure 3 is a schematic diagram of a refrigeration capacity prediction device based on transfer learning provided in an embodiment of the present application. Figure 3 As shown, the refrigeration capacity prediction device based on transfer learning includes: An acquisition module 301 is configured to acquire first historical data of a first location and second historical data of a second location, wherein the first historical data and the second historical data both include: location data, environment data, weather data and cooling capacity; A first training module 302 is configured to perform encoding and decoding training on the automatic encoding and decoding network using the first historical data, and perform cooling capacity prediction training on the prediction network; A freezing module 303 is configured to freeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; The second training module 304 is configured to perform encoding and decoding training on the automatic encoding and decoding network using the second historical data, and perform cooling capacity prediction training on the prediction network; The unfreezing module 305 is configured to unfreeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; The prediction module 306 is configured to use the automatic encoding and decoding network and the prediction network to predict the cooling capacity required by the second place at different times of the day.

[0035] According to the technical solution provided in the embodiment of the present application, the first historical data of the first place and the second historical data of the second place are obtained, wherein the first historical data and the second historical data both include: place data, environmental data, weather data and cooling capacity; the first historical data is used to perform encoding and decoding training on the automatic encoding and decoding network, and the prediction network is trained for cooling capacity prediction; the weights of the last layer of fully connected networks in the automatic encoding and decoding network and the prediction network are frozen respectively; the second historical data is used to perform encoding and decoding training on the automatic encoding and decoding network, and the prediction network is trained for cooling capacity prediction; the weights of the last layer of fully connected networks in the automatic encoding and decoding network and the prediction network are unfrozen respectively; the automatic encoding and decoding network and the prediction network are used to predict the cooling capacity required for the second place at different times of the day. The above-mentioned technical means can solve the problem in the prior art that the small amount of training data leads to overfitting of cooling capacity prediction and poor generalization ability, thereby avoiding overfitting of cooling capacity prediction and improving generalization ability.

[0036] In some embodiments, the first training module 302 is also configured to process the place data, environment data and weather data in the first historical data respectively through the automatic codec network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; use the mean square error loss function to calculate the reconstruction loss between the place data and the place reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data; optimize the weights in the automatic codec network based on the reconstruction loss to complete the codec training of the automatic codec network for the first historical data.

[0037] In some embodiments, the first training module 302 is further configured to, after using the first historical data to perform codec training on the automatic codec network, respectively process the place data, environment data and weather data in the first historical data through the automatic codec network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; process the place reconstruction data, environment reconstruction data and weather reconstruction data through the prediction network to obtain a cooling capacity prediction value corresponding to the first historical data; calculate the prediction loss between the cooling capacity prediction value and the cooling capacity in the first historical data using the regression loss function; and optimize the weights in the prediction network based on the prediction loss to complete the cooling capacity prediction training of the prediction network based on the first historical data.

[0038] In some embodiments, the second training module 304 is further configured to process the place data, environment data and weather data in the second historical data respectively through the automatic codec network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; use the mean square error loss function to calculate the reconstruction loss between the place data and the place reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data; optimize the weights in the automatic codec network based on the reconstruction loss to complete the codec training of the automatic codec network using the second historical data.

[0039] In some embodiments, the second training module 304 is further configured to, after using the second historical data to perform codec training on the automatic codec network, respectively process the location data, environment data and weather data in the second historical data through the automatic codec network to obtain location reconstruction data, environment reconstruction data and weather reconstruction data; process the location reconstruction data, environment reconstruction data and weather reconstruction data through the prediction network to obtain a cooling capacity prediction value corresponding to the second historical data; calculate the prediction loss between the cooling capacity prediction value and the cooling capacity in the second historical data using the regression loss function; and optimize the weights in the prediction network based on the prediction loss to complete the cooling capacity prediction training of the prediction network using the second historical data.

[0040] In some embodiments, the prediction module 306 is also configured to obtain the location data, environmental data, and weather data of the second place at the current moment; process the location data, environmental data, and weather data of the second place at the current moment respectively through the automatic encoding and decoding network to obtain location reconstruction data, environmental reconstruction data, and weather reconstruction data; process the location reconstruction data, environmental reconstruction data, and weather reconstruction data through the prediction network to obtain the cooling capacity prediction value of the second place at the current moment.

[0041] In some embodiments, the first training module 302 is further configured to construct an automatic encoding and decoding network using multiple layers of fully connected layers and activation functions; and to construct a prediction network using multiple layers of fully connected layers, batch normalization and activation functions.

[0042] In some embodiments, the first training module 302 is further configured to process the place data, environment data and weather data in the first historical data respectively through the automatic codec network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; use the mean square error loss function to respectively calculate the reconstruction loss between the place data and the place reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data; optimize the weights in the automatic codec network according to the reconstruction loss to complete the codec training of the automatic codec network by the first historical data; after the codec training of the automatic codec network by the first historical data, process the place data, environment data and weather data in the first historical data respectively through the automatic codec network to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; process the place reconstruction data, environment reconstruction data and weather reconstruction data through the prediction network to obtain the cooling capacity prediction value corresponding to the first historical data; use the regression loss function to calculate the prediction loss between the cooling capacity prediction value and the cooling capacity in the first historical data; optimize the weights in the prediction network according to the prediction loss to complete the cooling capacity prediction training of the prediction network by the first historical data.

[0043] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0044] Figure 4 Schematic diagram of an electronic device 4 provided in an embodiment of the present application. Figure 4 As shown, the electronic device 4 of this embodiment includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 401 executes the computer program 403, the functions of the modules / units in the above-mentioned device embodiments are implemented.

[0045] The electronic device 4 may be a desktop computer, a notebook, a PDA, a cloud server, or other electronic device. The electronic device 4 may include, but is not limited to, a processor 401 and a memory 402. Those skilled in the art will appreciate that Figure 4 The electronic device 4 is merely an example and does not limit the electronic device 4 , and may include more or less components than those shown in the figure, or different components.

[0046] The processor 401 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0047] The memory 402 may be an internal storage unit of the electronic device 4, for example, a hard disk or memory of the electronic device 4. The memory 402 may also be an external storage device of the electronic device 4, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 4. The memory 402 may also include both an internal storage unit and an external storage device of the electronic device 4. The memory 402 is used to store computer programs and other programs and data required by the electronic device.

[0048] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.

[0049] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. The computer program may include computer program code, and the computer program code may be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electric carrier signals and telecommunication signals.

[0050] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A refrigeration capacity prediction method based on transfer learning, characterized in that: include: Acquire first historical data of a first location and second historical data of a second location, wherein the first historical data and the second historical data both include: location data, environment data, weather data, and cooling capacity; Using the first historical data, the automatic encoding and decoding network is trained for encoding and decoding, and the prediction network is trained for refrigeration capacity prediction; Freeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; Using the second historical data, the automatic encoding and decoding network is trained for encoding and decoding, and the prediction network is trained for refrigeration capacity prediction; Unfreeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; The automatic encoding and decoding network and the prediction network are used to predict the cooling capacity required by the second place at different times of the day.

2. The method according to claim 1, characterized in that Performing encoding and decoding training on the automatic encoding and decoding network using the first historical data includes: Processing the location data, environment data and weather data in the first historical data respectively through the automatic encoding and decoding network to obtain location reconstruction data, environment reconstruction data and weather reconstruction data; Using a mean square error loss function, respectively calculating the reconstruction losses between the location data and the location reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data; The weights in the automatic encoding and decoding network are optimized according to the reconstruction loss to complete the encoding and decoding training of the automatic encoding and decoding network by the first historical data.

3. The method according to claim 1, characterized in that Using the first historical data to perform cooling capacity prediction training on the prediction network includes: After performing the encoding and decoding training on the automatic encoding and decoding network using the first historical data, the place data, the environment data and the weather data in the first historical data are processed by the automatic encoding and decoding network respectively to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; Processing the site reconstruction data, the environment reconstruction data and the weather reconstruction data through the prediction network to obtain a cooling capacity prediction value corresponding to the first historical data; Calculating a prediction loss between the predicted cooling capacity value and the cooling capacity in the first historical data using a regression loss function; The weights in the prediction network are optimized according to the prediction loss to complete the refrigeration capacity prediction training of the prediction network based on the first historical data.

4. The method according to claim 1, characterized in that: Performing the encoding and decoding training on the automatic encoding and decoding network using the second historical data includes: Processing the location data, environment data and weather data in the second historical data respectively through the automatic encoding and decoding network to obtain location reconstruction data, environment reconstruction data and weather reconstruction data; Using a mean square error loss function, respectively calculating the reconstruction losses between the location data and the location reconstruction data, the environment data and the environment reconstruction data, and the weather data and the weather reconstruction data; The weights in the automatic encoding and decoding network are optimized according to the reconstruction loss to complete the encoding and decoding training of the automatic encoding and decoding network by the second historical data.

5. The method according to claim 1, characterized in that Using the second historical data to perform cooling capacity prediction training on the prediction network includes: After performing the encoding and decoding training on the automatic encoding and decoding network using the second historical data, the place data, the environment data and the weather data in the second historical data are processed by the automatic encoding and decoding network respectively to obtain place reconstruction data, environment reconstruction data and weather reconstruction data; Processing the site reconstruction data, the environment reconstruction data and the weather reconstruction data through the prediction network to obtain a cooling capacity prediction value corresponding to the second historical data; Calculating a prediction loss between the predicted cooling capacity value and the cooling capacity in the second historical data using a regression loss function; The weights in the prediction network are optimized according to the prediction loss to complete the refrigeration capacity prediction training of the prediction network using the second historical data.

6. The method according to claim 1, characterized in that Predicting the cooling capacity required by the second place at different times of the day using the automatic encoding and decoding network and the prediction network includes: Acquire location data, environment data, and weather data of the second location at the current moment; Processing the location data, environment data, and weather data of the second location at the current moment respectively through the automatic encoding and decoding network to obtain location reconstruction data, environment reconstruction data, and weather reconstruction data; The location reconstruction data, the environment reconstruction data and the weather reconstruction data are processed by the prediction network to obtain a cooling capacity prediction value of the second location at the current moment.

7. The method according to claim 1, characterized in that Before performing encoding and decoding training on the automatic encoding and decoding network using the first historical data and performing cooling capacity prediction training on the prediction network, the method further includes: The automatic encoding and decoding network is constructed using multiple layers of fully connected layers and activation functions; The prediction network is constructed using multiple layers of fully connected layers, batch normalization, and activation functions.

8. A refrigeration capacity prediction device based on transfer learning, characterized in that: include: An acquisition module is configured to acquire first historical data of a first location and second historical data of a second location, wherein the first historical data and the second historical data both include: location data, environment data, weather data and cooling capacity; A first training module is configured to perform encoding and decoding training on the automatic encoding and decoding network using the first historical data, and perform cooling capacity prediction training on the prediction network; A freezing module, configured to freeze the weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; a second training module, configured to perform the encoding and decoding training on the automatic encoding and decoding network using the second historical data, and to perform the refrigeration capacity prediction training on the prediction network; A thawing module, configured to thawing weights of the last fully connected network layer in the automatic encoding and decoding network and the prediction network respectively; The prediction module is configured to use the automatic encoding and decoding network and the prediction network to predict the cooling capacity required by the second place at different times of the day.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images