Predictable method for terminals and electronic device

Abstract

Embodiments of the present application provide a terminal energy consumption prediction method and electronic equipment, and relate to the field of energy consumption planning, comprising: obtaining terminal energy consumption historical data, and obtaining at least two terminal energy consumption sequences without repeated data by binning the terminal energy consumption historical data; performing data coding operation on the terminal energy consumption sequences to obtain terminal energy consumption sequence features; for each terminal energy consumption sequence feature, obtaining a first correlation index between terminal energy consumption sequences matched with the terminal energy consumption sequence feature and terminal energy consumption sequence features matched with associated terminal energy consumption sequence features of the terminal energy consumption sequence feature, and determining a first weight index corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the first correlation index; determining a first attention feature in all terminal energy consumption sequence features based on the first weight index; and performing terminal energy consumption prediction based on the first attention feature to obtain terminal energy consumption prediction information. The effect of terminal energy consumption prediction can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of energy, and more specifically, to a method and electronic device for predicting end-user energy consumption. Background Technology

[0002] Current research on end-user energy service demand forecasting, both domestically and internationally, mainly focuses on identifying typical load characteristics from electricity loads and classifying terminals based on their attribute information, thereby providing personalized energy services for different types of terminals.

[0003] In energy consumption monitoring, predictive methods can be used to forecast future energy consumption to provide data references. Energy consumption data is typically treated as time-series data, and the energy consumption trends within it are used to predict energy consumption within a target time period. However, current methods for predicting end-user energy consumption suffer from low accuracy due to the sheer volume of data. Therefore, identifying the most representative characteristic data from this vast amount of end-user energy consumption data is a pressing issue that needs to be addressed. Summary of the Invention

[0004] The purpose of this invention is to provide a method and electronic device for predicting terminal power consumption.

[0005] According to a first aspect of the present invention, a method for predicting terminal power consumption is provided, comprising:

[0006] Acquire historical energy consumption data from terminals, and divide the historical energy consumption data into bins to obtain at least two terminal energy consumption sequences that do not contain duplicate data;

[0007] Perform data encoding operations on at least two of the terminal power consumption sequences to obtain at least two terminal power consumption sequence features;

[0008] For each of the terminal energy consumption sequence features, a first correlation index is obtained between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature. Based on the terminal energy consumption sequence feature and the first correlation index, a first weight index corresponding to the terminal energy consumption sequence feature is determined.

[0009] Based on the first weighting index, at least two first features of interest are determined from all the terminal energy consumption sequence features;

[0010] Terminal energy consumption prediction is performed based on at least two of the first features of interest to obtain terminal energy consumption prediction information.

[0011] Optionally, determining the first weight index corresponding to the terminal energy consumption sequence characteristics based on the terminal energy consumption sequence characteristics and the first correlation index includes:

[0012] Based on the terminal energy consumption sequence characteristics and the first correlation index, the bin attention characteristics corresponding to the terminal energy consumption sequence characteristics are determined;

[0013] Based on the bin-specific attention characteristics and the first correlation index, a first weight index corresponding to the terminal energy consumption sequence characteristics is determined;

[0014] The step of determining the first weight index corresponding to the terminal energy consumption sequence characteristics based on the binning attention characteristics and the first correlation index includes:

[0015] The binning attention features are recoded into the weight index domain.

[0016] The weight index features corresponding to the terminal energy consumption sequence features are obtained;

[0017] Based on the weight index features and the first correlation index, the first weight index corresponding to the terminal energy consumption sequence features is determined.

[0018] Optionally, the number of associated terminal energy consumption sequence features is at least two; determining the first weight index corresponding to the terminal energy consumption sequence feature based on the weight index feature and the first correlation index includes:

[0019] Based on the weight index features and the first correlation index corresponding to at least two associated terminal energy consumption sequence features, perform product and sum operations respectively to obtain the first weight index corresponding to the terminal energy consumption sequence features.

[0020] Optionally, determining the bin-specific attention characteristics corresponding to the terminal energy consumption sequence characteristics based on the terminal energy consumption sequence characteristics and the first correlation index includes:

[0021] Based on the terminal energy consumption sequence characteristics and the first correlation index, a feature attention statistics operation is performed to obtain the bin attention index corresponding to the terminal energy consumption sequence characteristics.

[0022] Based on the terminal energy consumption sequence characteristics and the sub-container attention index, the sub-container attention characteristics corresponding to the terminal energy consumption sequence characteristics are determined;

[0023] The step of performing feature attention statistics based on the terminal energy consumption sequence characteristics and the first correlation index to obtain the bin-specific attention index corresponding to the terminal energy consumption sequence characteristics includes:

[0024] Based on the terminal energy consumption sequence characteristics, determine the energy consumption characteristics and time characteristics corresponding to the terminal energy consumption sequence characteristics;

[0025] Based on the energy consumption characteristics, the time characteristics, and the first correlation index, the bin attention index corresponding to the terminal energy consumption sequence characteristics is determined.

[0026] Optionally, determining the bin attention feature corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the bin attention index includes:

[0027] Based on the bin-specific attention index, a weighted summation operation is performed on the terminal energy consumption sequence features to obtain the bin-specific attention features corresponding to the terminal energy consumption sequence features.

[0028] Optionally, obtaining a first correlation index between the terminal energy consumption sequence that matches the terminal energy consumption sequence features and the terminal energy consumption sequence that matches the associated terminal energy consumption sequence features includes:

[0029] Obtain a first time period in the historical data of terminal energy consumption that matches the terminal energy consumption sequence features, and a second time period in the historical data of terminal energy consumption that matches the associated terminal energy consumption sequence features.

[0030] Based on the first time period and the second time period, determine the time-related information corresponding to the terminal energy consumption sequence characteristics;

[0031] The time-related information is used as the first correlation index between the terminal energy consumption sequence that matches the terminal energy consumption sequence feature and the terminal energy consumption sequence that matches the associated terminal energy consumption sequence feature.

[0032] Optionally, determining the time-related information corresponding to the terminal energy consumption sequence characteristics based on the first time period and the second time period includes:

[0033] Determine the time offset parameter between the first time period and the second time period;

[0034] A time-related encoding operation is performed on the time offset parameter to obtain the time-related information corresponding to the terminal energy consumption sequence characteristics.

[0035] Optionally, the step of performing terminal energy consumption prediction based on at least two of the first features of interest to obtain terminal energy consumption prediction information includes:

[0036] Perform multiple rounds of attention feature separation operation based on at least two first attention features, and obtain at least two second attention features obtained when performing the attention feature separation operation in each round;

[0037] Perform a feature combination operation on the first feature of interest and all the second features of interest to obtain the energy-useful joint features;

[0038] Based on the aforementioned joint energy consumption characteristics, terminal energy consumption prediction is performed to obtain terminal energy consumption prediction information.

[0039] Optionally, the feature separation operation performed in each round includes:

[0040] Obtain at least two target energy consumption features, wherein the target energy consumption feature is the first feature of interest, or the second feature of interest obtained when the feature of interest separation operation was performed in the previous round;

[0041] For each of the target energy consumption characteristics, a second correlation index is obtained between the terminal energy consumption sequence that matches the target energy consumption characteristic and the terminal energy consumption sequence that matches the associated energy consumption characteristic of the target energy consumption characteristic. Based on the target energy consumption characteristic and the second correlation index, a second weight index corresponding to the target energy consumption characteristic is determined.

[0042] Based on the second weighting index, at least two second features of interest are determined from all the target energy consumption features.

[0043] According to a first aspect of the present invention, an electronic device is provided, comprising:

[0044] A storage device on which computer programs are stored;

[0045] A processing device for executing the computer program in the storage device to implement the steps of the method described in the first aspect.

[0046] The beneficial effects of this invention include, for example: after obtaining historical terminal energy consumption data, the historical terminal energy consumption data is first binned to obtain at least two terminal energy consumption sequences that do not contain duplicate data, and data encoding operations are performed on the at least two terminal energy consumption sequences to obtain at least two terminal energy consumption sequence features; by performing binning operations on the historical terminal energy consumption data to obtain at least two terminal energy consumption sequences that do not contain duplicate data, the encoding efficiency of the terminal energy consumption sequence features of these terminal energy consumption sequences can be improved; in addition, after obtaining at least two terminal energy consumption sequence features, for each terminal energy consumption sequence feature, the first correlation between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature is first obtained. The system firstly identifies the first weight index corresponding to the terminal energy consumption sequence characteristics based on the terminal energy consumption sequence characteristics and the first correlation index. Since the first correlation index is the correlation index between the terminal energy consumption sequence characteristics that match the terminal energy consumption sequence characteristics and the terminal energy consumption sequence characteristics that match the associated terminal energy consumption sequence characteristics, the first correlation index can characterize the correlation information between the terminal energy consumption sequence characteristics and the associated terminal energy consumption sequence characteristics. Therefore, when determining the first weight index corresponding to the terminal energy consumption sequence characteristics based on the terminal energy consumption sequence characteristics and the first correlation index, the terminal energy consumption sequence characteristics and their corresponding correlation information can be combined, so that the determined first weight index can more accurately express the importance of the terminal energy consumption sequence characteristics compared to the associated terminal energy consumption sequence characteristics. After determining the first weight index corresponding to each terminal energy consumption sequence feature, at least two first features of interest are first identified from all terminal energy consumption sequence features based on the first weight index. Then, terminal energy consumption prediction is performed based on the at least two first features of interest to obtain terminal energy consumption prediction information. Through the first weight index, at least two first features of interest that can better express the historical data of terminal energy consumption can be identified from all terminal energy consumption sequence features, thereby improving the feature extraction effect of the historical data of terminal energy consumption. This can improve the accuracy of terminal energy consumption prediction when performing terminal energy consumption prediction based on these first features of interest, and thus improve the effect of terminal energy consumption prediction. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart of a terminal energy consumption prediction method provided in an embodiment of this application.

[0049] Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0051] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0052] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0053] In the description of this invention, it should be noted that if terms such as "upper," "lower," "inner," or "outer" are used to indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of this invention is usually placed, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0054] Furthermore, the terms "first" and "second" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0055] It should be noted that, where there is no conflict, the features in the embodiments of the present invention can be combined with each other.

[0056] Figure 1 This is a flowchart illustrating a method for predicting terminal power consumption according to an embodiment of this application. This method can be executed by a server, a terminal device, or any electronic device with information processing capabilities. In this embodiment, the method is described using a server as an example. (Refer to...) Figure 1 The terminal uses predictive methods including, but not limited to, steps 110 to 150.

[0057] Step S110: Obtain historical terminal energy consumption data, and divide the historical terminal energy consumption data into bins to obtain at least two terminal energy consumption sequences that do not contain duplicate data.

[0058] In one embodiment, when the terminal's data acquisition device collects historical energy consumption data of the terminal, the server can directly obtain the historical energy consumption data uploaded by the terminal. This historical energy consumption data can be stored in the form of energy consumption sequences, for example, the data may include energy consumption and the corresponding time. Then, the historical energy consumption data is binned to obtain at least two terminal energy consumption sequences that do not contain duplicate data. By performing a binning operation on the historical energy consumption data to obtain at least two smaller time-range energy consumption data, i.e., terminal energy consumption sequences, the data within the entire time range can be binned to obtain at least two smaller time-range energy consumption data. This ensures that the content of the binned terminal energy consumption sequences is mainly based on terminal energy consumption information within that time period. Therefore, when performing data encoding operations on these terminal energy consumption sequences, the influence of energy consumption data from other time periods on the encoding can be reduced. This not only improves the extraction efficiency of terminal energy consumption sequence features but also improves the extraction accuracy of these terminal energy consumption sequence features.

[0059] In one embodiment, when dividing the terminal energy consumption history data into bins to obtain at least two terminal energy consumption sequences that do not contain duplicate data, the number of terminal energy consumption sequences to be binned can be determined first. Then, the length of the terminal energy consumption sequence can be determined based on the number of terminal energy consumption sequences and the sequence length of the terminal energy consumption history data. Next, the terminal energy consumption history data can be binned based on the length of the terminal energy consumption sequence to obtain at least two terminal energy consumption sequences that do not contain duplicate data.

[0060] Step S120: Perform data encoding operations on at least two terminal power consumption sequences to obtain at least two terminal power consumption sequence features.

[0061] In one embodiment, after obtaining at least two terminal power consumption sequences that do not contain duplicate data through binning, data encoding operations can be performed on these at least two terminal power consumption sequences to obtain at least two terminal power consumption sequence features. That is, data encoding operations can be performed on each terminal power consumption sequence to obtain the terminal power consumption sequence features of each terminal power consumption sequence.

[0062] In one embodiment, the at least two terminal power sequence features obtained by performing data encoding operations on at least two terminal power sequences can be represented in the form of feature vectors. Therefore, after performing data encoding operations on at least two terminal power sequences, a terminal power sequence feature vector composed of at least two terminal power sequence features can be obtained, so that these terminal power sequence features can correspond to the terminal power sequence vectors obtained in the previous steps, thereby facilitating subsequent steps to process the terminal power sequence features in the form of feature vectors.

[0063] Step S130: For each terminal energy consumption sequence feature, obtain the first correlation index between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature, and determine the first weight index corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the first correlation index.

[0064] In one embodiment, after extracting the terminal energy consumption sequence features of each terminal energy consumption sequence, for each terminal energy consumption sequence feature, a first correlation index can be obtained between its matching terminal energy consumption sequence and the terminal energy consumption sequence matching its associated terminal energy consumption sequence feature. Since the first correlation index is the correlation index between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature, the first correlation index can characterize the information of the association between the terminal energy consumption sequence feature and the associated terminal energy consumption sequence feature, such as context information. Therefore, when determining the first weight index corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the first correlation index, the terminal energy consumption sequence feature can be combined with its corresponding context information, so that the determined first weight index can more accurately indicate the weight of the terminal energy consumption sequence feature relative to the associated terminal energy consumption sequence feature.

[0065] In one embodiment, in the process of obtaining the first correlation index between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the associated terminal energy consumption sequence feature matching the terminal energy consumption sequence feature, the process can first obtain the first time period of the terminal energy consumption sequence matching the terminal energy consumption sequence feature in the terminal energy consumption historical data, and the second time period of the associated terminal energy consumption sequence feature matching the terminal energy consumption sequence feature in the terminal energy consumption historical data. Then, based on the first time period and the second time period, the time-related information corresponding to the terminal energy consumption sequence feature is determined. Subsequently, the time-related information is used as the first correlation index between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the associated terminal energy consumption sequence feature matching the terminal energy consumption sequence feature. Since the first correlation index can characterize the correlation information between terminal energy consumption sequence features and associated terminal energy consumption sequence features, such as contextual information, the time-related information determined based on the first and second time periods can characterize the contextual information between terminal energy consumption sequences and adjacent terminal energy consumption sequences. Since terminal energy consumption sequence features correspond one-to-one with terminal energy consumption sequences, the time-related information matching a terminal energy consumption sequence feature can be represented using the time-related information of its matching terminal energy consumption sequence. Therefore, the time-related information corresponding to a terminal energy consumption sequence feature can be used as the first correlation index between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature. Furthermore, by first obtaining the first time period of the terminal energy consumption sequence matching the terminal energy consumption sequence feature in the historical terminal energy consumption data and the second time period of the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature in the historical terminal energy consumption data, and then determining the time-related information used as the first correlation index based on the first and second time periods, the computational load of the time-related information matching the terminal energy consumption sequence feature can be effectively reduced, making the determination of the time-related information matching the terminal energy consumption sequence feature more convenient.

[0066] In one embodiment, in determining the time-related information corresponding to the terminal energy consumption sequence feature based on a first time period and a second time period, a time offset parameter between the first and second time periods can be determined first. Then, a time relationship encoding operation is performed on the time offset parameter to obtain the time-related information corresponding to the terminal energy consumption sequence feature. Specifically, when performing the time relationship encoding operation on the time offset parameter, a time weight for performing the time relationship encoding operation can be determined first, and then the time relationship encoding operation is performed on the time offset parameter based on the time weight. In one embodiment, when it is necessary to determine the time-related information corresponding to the terminal energy consumption sequence feature, the first and second time periods corresponding to the terminal energy consumption sequence feature are first obtained. Then, the time offset parameter between the first and second time periods is determined. Next, the time offset parameter is input to a fully connected layer, which assigns a time weight to the time offset parameter. The time-related information corresponding to the terminal energy consumption sequence feature is obtained by multiplying the time weight by the time offset parameter.

[0067] In one embodiment, in determining the first weight index corresponding to a terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the first correlation index, the binning attention feature corresponding to the terminal energy consumption sequence feature can be determined first based on the terminal energy consumption sequence feature and the first correlation index. Then, the first weight index corresponding to the terminal energy consumption sequence feature can be determined based on the binning attention feature and the first correlation index. Since the first correlation index can characterize the correlation information between the terminal energy consumption sequence feature and related terminal energy consumption sequence features, and the binning attention feature corresponding to the terminal energy consumption sequence feature determined based on the terminal energy consumption sequence feature and the first correlation index can highlight the weight of the terminal energy consumption sequence feature relative to the related terminal energy consumption sequence features, the first weight index corresponding to the terminal energy consumption sequence feature determined based on the binning attention feature and the first correlation index can better express the importance of each terminal energy consumption sequence feature in all terminal energy consumption sequence features. This is beneficial for subsequent steps to determine multiple more important first attention features that can better express the features in the energy consumption data based on the first weight index in all terminal energy consumption sequence features, thereby improving the accuracy of the subsequent terminal energy consumption prediction.

[0068] In one embodiment, during the process of determining the bin-specific attention features corresponding to the terminal energy consumption sequence features based on the terminal energy consumption sequence features and the first correlation index, a feature attention statistics operation can first be performed based on the terminal energy consumption sequence features and the first correlation index to obtain the bin-specific attention index corresponding to the terminal energy consumption sequence features. Then, based on the terminal energy consumption sequence features and the bin-specific attention index, the bin-specific attention features corresponding to the terminal energy consumption sequence features are determined. By determining the bin-specific attention index corresponding to the terminal energy consumption sequence features based on the terminal energy consumption sequence features and the first correlation index, the bin-specific attention index can express the attention level of the terminal energy consumption sequence features. This allows the bin-specific attention features determined based on the terminal energy consumption sequence features and the bin-specific attention index to enhance the terminal energy consumption sequence features, thereby improving the prediction effect of subsequent steps on terminal energy consumption.

[0069] In one embodiment, during the process of performing feature attention statistics based on terminal energy consumption sequence characteristics and a first correlation index to obtain the bin attention index corresponding to the terminal energy consumption sequence characteristics, the energy consumption characteristics and time characteristics corresponding to the terminal energy consumption sequence characteristics can be determined first based on the terminal energy consumption sequence characteristics. Then, based on the energy consumption characteristics, time characteristics, and the first correlation index, the bin attention index corresponding to the terminal energy consumption sequence characteristics can be determined. Specifically, when determining the energy consumption characteristics and time characteristics corresponding to the terminal energy consumption sequence characteristics, two different one-dimensional mappings or linear transformations can be performed on the terminal energy consumption sequence characteristics to obtain the energy consumption characteristics and time characteristics corresponding to the terminal energy consumption sequence characteristics. In one embodiment, when it is necessary to determine the binning attention index of terminal energy consumption sequence features, a one-dimensional mapping of energy consumption features can be performed on the terminal energy consumption sequence features to obtain the energy consumption features corresponding to the terminal energy consumption sequence features, and a one-dimensional mapping of time features can be performed on the terminal energy consumption sequence features to obtain the time features corresponding to the terminal energy consumption sequence features, and the time-related information corresponding to the terminal energy consumption sequence features can be determined. Then, the time features corresponding to the terminal energy consumption sequence features are transposed to obtain the transposed time features. Then, the result of vector product of energy consumption features and transposed time features is vector summed with the time-related information. Finally, the result of vector summation is input into a preset function to perform probability distribution mapping to obtain the binning attention index of terminal energy consumption sequence features.

[0070] In one embodiment, in the process of determining the bin attention feature corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the bin attention index, a weighted summation operation can be performed on the terminal energy consumption sequence feature based on the bin attention index to obtain the bin attention feature corresponding to the terminal energy consumption sequence feature.

[0071] In one embodiment, during the process of determining the first weight index corresponding to the terminal energy consumption sequence feature based on the binning attention feature and the first correlation index, the binning attention feature can be re-encoded into the weight index domain to obtain the weight index feature corresponding to the terminal energy consumption sequence feature. Then, based on the weight index feature and the first correlation index, the first weight index corresponding to the terminal energy consumption sequence feature is determined. Specifically, when re-encoding the binning attention feature into the weight index domain, a fully connected layer can be used. Furthermore, when there are at least two associated terminal energy consumption sequence features, during the process of determining the first weight index corresponding to the terminal energy consumption sequence feature based on the weight index feature and the first correlation index, a product and sum operation can be performed on the weight index feature and the first correlation index corresponding to the at least two associated terminal energy consumption sequence features to obtain the first weight index corresponding to the terminal energy consumption sequence feature. In one embodiment, when it is necessary to determine the first weight index corresponding to the terminal energy consumption sequence feature, at least two time-related information that match the terminal energy consumption sequence feature can be determined first. Then, the binning attention feature is re-encoded into the weight index domain using a preset encoder to obtain the weight index feature. The weight index feature is then normalized using a preset function. Next, the normalized weight index feature is multiplied and accumulated with at least two time-related information to obtain the first weight index corresponding to the terminal energy consumption sequence feature.

[0072] Step S140: Based on the first weight index, determine at least two first features of interest from all terminal energy consumption sequence features.

[0073] In one embodiment, during the process of determining at least two first features of interest from all terminal energy consumption sequence features based on a first weight index, at least two target weight indices that meet preset conditions can be determined firstly from all first weight indices. Then, from all terminal energy consumption sequence features, at least two terminal energy consumption sequence features that match the at least two target weight indices are determined as at least two first features of interest. Although the first weight index can express the importance of terminal energy consumption sequence features, a threshold index is still needed to distinguish the first weight indices that can express the first features of interest. Therefore, a threshold index can be preset, and cases where the weight index is greater than the threshold index can be used as preset conditions. At least two target weight indices that meet the preset conditions can be determined from all first weight indices. At this point, terminal energy consumption sequence features that match these target weight indices can be determined as first features of interest from all terminal energy consumption sequence features.

[0074] Step S150: Perform terminal energy consumption prediction based on at least two first-interest features to obtain terminal energy consumption prediction information.

[0075] In one embodiment, after obtaining at least two first features of interest, terminal power consumption prediction can be performed based on these at least two first features of interest to obtain terminal power consumption prediction information. Specifically, in the process of performing terminal power consumption prediction based on the at least two first features of interest to obtain feature vectors of interest, the at least two first features of interest can first be vectorized to obtain feature vectors of interest. Then, a terminal power consumption prediction model is used to perform terminal power consumption prediction on the feature vectors of interest to obtain terminal power consumption prediction information. When performing vectorization on the at least two first features of interest, a multilayer perceptron can be used to perform vectorization on the at least two first features of interest.

[0076] In one embodiment, before performing terminal energy consumption prediction on the feature vector of interest using the terminal energy consumption prediction model, the terminal energy consumption prediction model can be trained in advance. For example, the method described in the previous embodiment can be used to obtain at least two first features of interest in the training samples, and these first features of interest in the training samples can be vectorized to obtain the feature vector of interest in the training samples. Then, the feature vector of interest in the training samples is input into the terminal energy consumption prediction model to perform terminal energy consumption prediction and obtain the prediction result. Then, based on the prediction result and the sample label, the prediction loss value is determined, and gradient backpropagation is performed in the terminal energy consumption prediction model based on the prediction loss value to correct the model parameters of the terminal energy consumption prediction model, thereby realizing the training of the terminal energy consumption prediction model.

[0077] In this embodiment, the terminal energy consumption prediction method, including steps 110 to 150 above, after obtaining historical terminal energy consumption data, firstly bins the historical terminal energy consumption data to obtain at least two terminal energy consumption sequences that do not contain duplicate data, and then performs data encoding operations on the at least two terminal energy consumption sequences to obtain at least two terminal energy consumption sequence features. By performing binning operations on the historical terminal energy consumption data to obtain at least two terminal energy consumption sequences that do not contain duplicate data, the encoding efficiency of the terminal energy consumption sequence features of these terminal energy consumption sequences can be improved. In addition, after obtaining at least two terminal energy consumption sequence features, for each terminal energy consumption sequence feature, firstly, the terminal energy consumption sequence matching the terminal energy consumption sequence feature is obtained, and then the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature is obtained. The first correlation index between columns is used, and then the first weight index corresponding to the terminal energy consumption sequence feature is determined based on the terminal energy consumption sequence feature and the first correlation index. Since the first correlation index is the correlation index between the terminal energy consumption sequence feature that matches the terminal energy consumption sequence feature and the terminal energy consumption sequence feature that matches the associated terminal energy consumption sequence feature, the first correlation index can characterize the correlation information between the terminal energy consumption sequence feature and the associated terminal energy consumption sequence feature. Therefore, when determining the first weight index corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the first correlation index, the terminal energy consumption sequence feature and its corresponding correlation information can be combined, so that the determined first weight index can more accurately express the importance of the terminal energy consumption sequence feature compared to the associated terminal energy consumption sequence feature. After determining the first weight index corresponding to each terminal energy consumption sequence feature, at least two first features of interest are first identified from all terminal energy consumption sequence features based on the first weight index. Then, terminal energy consumption prediction is performed based on the at least two first features of interest to obtain terminal energy consumption prediction information. Through the first weight index, at least two first features of interest that can better express the historical data of terminal energy consumption can be identified from all terminal energy consumption sequence features, thereby improving the feature extraction effect of the historical data of terminal energy consumption. This can improve the accuracy of terminal energy consumption prediction when performing terminal energy consumption prediction based on these first features of interest, and thus improve the effect of terminal energy consumption prediction.

[0078] In one embodiment, during the process of performing terminal energy consumption prediction based on at least two first features of interest to obtain terminal energy consumption prediction information, multiple rounds of feature separation operations can be performed based on at least two first features of interest to obtain at least two second features of interest obtained in each round of feature separation operations. Then, a feature combination operation is performed on the first features of interest and all second features of interest to obtain joint energy consumption features. Finally, terminal energy consumption prediction is performed based on the joint energy consumption features to obtain terminal energy consumption prediction information. In this embodiment, the process of performing terminal energy consumption prediction based on the joint energy consumption features can be referred to the relevant description of step 150 above, and will not be repeated here.

[0079] In one embodiment, after obtaining at least two first features of interest, these first features of interest can be used as the basis to perform multiple rounds of feature separation operations, and then at least two second features of interest obtained in each round of feature separation operations can be acquired. Specifically, performing multiple rounds of feature separation operations based on these first features of interest means first performing a first round of feature separation operations on these first features of interest to obtain at least two second features of interest output by the first round of feature separation operations; then performing a second round of feature separation operations on these second features of interest output by the first round of feature separation operations to obtain at least two second features of interest output by the second round of feature separation operations; then performing a third round of feature separation operations on these second features of interest output by the second round of feature separation operations to obtain at least two second features of interest output by the third round of feature separation operations; and so on iteratively until the number of feature separation operations performed reaches a preset number. It should be noted that the preset number of operations can be appropriately selected based on the actual application situation; for example, the preset number of operations could be 3, 5, or 7 times, etc., and is not specifically limited here. By performing multiple rounds of feature separation operations based on these first features of interest, more accurate extraction of second features of interest can be performed on these first features of interest. This allows the obtained second features of interest to more accurately represent the features of terminal energy consumption, thereby improving the extraction effect of terminal energy consumption features.

[0080] In one embodiment, the feature separation operation performed in each round may include the following steps:

[0081] At least two target energy consumption features are obtained, wherein the target energy consumption feature is either the first feature of interest or the second feature of interest obtained during the previous round of feature separation operation; for each target energy consumption feature, a second correlation index is obtained between the terminal energy consumption sequence matching the target energy consumption feature and the terminal energy consumption sequence matching the associated energy consumption feature of the target energy consumption feature, and a second weight index is determined based on the target energy consumption feature and the second correlation index; based on the second weight index, at least two second features of interest are determined among all target energy consumption features.

[0082] Specifically, acquiring at least two target energy consumption features during the first round of feature separation operation can refer to acquiring at least two first-focused features. Acquiring at least two target energy consumption features during non-first-round feature separation operations refers to acquiring at least two second-focused features obtained in the previous round of feature separation operation. Therefore, performing feature separation operations multiple times constitutes a multi-round feature separation operation, allowing the obtained second-focused features to more accurately represent the terminal's energy consumption features. Furthermore, based on the steps included in the feature separation operation, steps 130 to 140 described above are essentially the content of the feature separation operation. Therefore, the process of acquiring the second correlation index between the terminal energy consumption sequence matching the target energy consumption feature and the terminal energy consumption sequence matching the associated energy consumption feature of the target energy consumption feature, the process of determining the second weight index corresponding to the target energy consumption feature based on the target energy consumption feature and the second correlation index, and the process of determining at least two second-focused features among all target energy consumption features based on the second weight index can all refer to the relevant descriptions in the previous embodiments, and will not be repeated here. In one embodiment, after obtaining at least two second features of interest obtained during the feature separation operation in each round, these second features of interest and the first features of interest can be combined to obtain joint energy consumption features. This joint energy consumption features can improve the accuracy of the expression of terminal energy consumption features in the historical data of terminal energy consumption, thereby improving the prediction effect of terminal energy consumption.

[0083] In one embodiment, when performing a feature combination operation on the first feature of interest and all second features of interest to obtain a combined energy-use feature, the first feature of interest and the second features of interest obtained during each round of feature separation operations can be combined to obtain the combined energy-use feature. To ensure the order information of the first feature of interest and the second features of interest obtained during each feature separation operation, the first feature of interest and the second features of interest obtained during each feature separation operation can be combined according to the order in which the feature separation operations are performed.

[0084] In one embodiment, during the process of extracting features from historical terminal energy consumption data using a terminal energy consumption prediction model to obtain terminal energy consumption feature vectors, a three-iteration feature separation operation can be performed based on the historical terminal energy consumption data. Specifically, when extracting features from historical terminal energy consumption data using the terminal energy consumption prediction model, the historical terminal energy consumption data can first be binned to obtain at least two terminal energy consumption sequences that do not contain duplicate data. Data encoding is then performed on these terminal energy consumption sequences to obtain at least two terminal energy consumption sequence features. Then, a first round of feature separation is performed on these terminal energy consumption sequence features. During this first round of feature separation, the binning attention index for each terminal energy consumption sequence feature is first determined. Based on the binning attention index of each terminal energy consumption sequence feature, feature adjustment is performed on each terminal energy consumption sequence feature to obtain each terminal energy consumption sequence. The process involves identifying binning attention features, determining the binning attention corresponding to each binning attention feature, and then selecting the top K (K ≥ 1) binning attention features with matching values ​​from largest to smallest. These K features are identified as at least two attention features obtained during the first round of attention feature separation. Next, feature combination is performed on these obtained attention features, followed by a second round of attention feature separation, until joint energy consumption features are obtained. Finally, the terminal energy consumption prediction model is invoked, inputting the terminal energy consumption feature vector into the model so that the feature recognition model within the model can determine the terminal energy consumption prediction information. This effectively improves the reliability of terminal energy consumption prediction.

[0085] The following is for reference. Figure 2 This document illustrates a structural schematic of an electronic device (e.g., a server or terminal device) 200 suitable for implementing embodiments of the present disclosure. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 2 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0086] like Figure 2As shown, electronic device 200 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 201, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 202 or a program loaded from storage device 208 into random access memory (RAM) 203. The RAM 203 also stores various programs and data required for the operation of electronic device 200. The processing device 201, ROM 202, and RAM 203 are interconnected via bus 204. Input / output (I / O) interface 205 is also connected to bus 204.

[0087] Typically, the following devices can be connected to I / O interface 205: input devices 206 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 207 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 208 including, for example, magnetic tapes, hard disks, etc.; and communication devices 209. Communication device 209 allows electronic device 200 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 2 An electronic device 200 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0088] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 209, or installed from storage device 208, or installed from ROM 202. When the computer program is executed by processing device 201, it performs the functions defined in the methods of embodiments of this disclosure.

[0089] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0090] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0091] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0092] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform one or more steps of the aforementioned method.

[0093] Alternatively, the aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to perform one or more steps of the aforementioned method.

[0094] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0095] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0096] The modules described in the embodiments of this disclosure can be implemented in software or hardware. The names of the modules are not, in some cases, intended to limit the functionality of the module itself.

[0097] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0098] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0099] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for predicting terminal energy consumption, characterized in that, include: Acquire historical energy consumption data from terminals, and divide the historical energy consumption data into bins to obtain at least two terminal energy consumption sequences that do not contain duplicate data; Perform data encoding operations on at least two of the terminal power consumption sequences to obtain at least two terminal power consumption sequence features; For each of the terminal energy consumption sequence features, a first correlation index is obtained between the terminal energy consumption sequence matching the terminal energy consumption sequence feature and the terminal energy consumption sequence matching the associated terminal energy consumption sequence feature. Based on the terminal energy consumption sequence feature and the first correlation index, a first weight index corresponding to the terminal energy consumption sequence feature is determined. Based on the first weighting index, at least two first features of interest are determined from all the terminal energy consumption sequence features; Based on at least two of the first features of interest, terminal energy consumption prediction is performed to obtain terminal energy consumption prediction information; The step of determining the first weight index corresponding to the terminal energy consumption sequence characteristics based on the terminal energy consumption sequence characteristics and the first correlation index includes: Based on the terminal energy consumption sequence characteristics and the first correlation index, the bin attention characteristics corresponding to the terminal energy consumption sequence characteristics are determined; Based on the bin-specific attention characteristics and the first correlation index, a first weight index corresponding to the terminal energy consumption sequence characteristics is determined; The step of determining the first weight index corresponding to the terminal energy consumption sequence characteristics based on the binning attention characteristics and the first correlation index includes: The binning attention features are recoded into the weight index domain. The weight index features corresponding to the terminal energy consumption sequence features are obtained; Based on the weight index features and the first correlation index, the first weight index corresponding to the terminal energy consumption sequence features is determined. The first correlation index obtained between the terminal energy consumption sequence that matches the terminal energy consumption sequence features and the terminal energy consumption sequence that matches the associated terminal energy consumption sequence features includes: Obtain a first time period in the historical data of terminal energy consumption that matches the terminal energy consumption sequence features, and a second time period in the historical data of terminal energy consumption that matches the associated terminal energy consumption sequence features. Based on the first time period and the second time period, determine the time-related information corresponding to the terminal energy consumption sequence characteristics; The time-related information is used as the first correlation index between the terminal energy consumption sequence that matches the terminal energy consumption sequence feature and the terminal energy consumption sequence that matches the associated terminal energy consumption sequence feature.

2. The method according to claim 1, characterized in that, The number of associated terminal energy consumption sequence features is at least two; determining the first weight index corresponding to the terminal energy consumption sequence feature based on the weight index feature and the first correlation index includes: Based on the weight index features and the first correlation index corresponding to at least two associated terminal energy consumption sequence features, perform product and sum operations respectively to obtain the first weight index corresponding to the terminal energy consumption sequence features.

3. The method according to claim 1, characterized in that, The step of determining the bin-specific attention characteristics corresponding to the terminal energy consumption sequence characteristics based on the terminal energy consumption sequence characteristics and the first correlation index includes: Based on the terminal energy consumption sequence characteristics and the first correlation index, a feature attention statistics operation is performed to obtain the bin attention index corresponding to the terminal energy consumption sequence characteristics. Based on the terminal energy consumption sequence characteristics and the sub-container attention index, the sub-container attention characteristics corresponding to the terminal energy consumption sequence characteristics are determined; The step of performing feature attention statistics based on the terminal energy consumption sequence characteristics and the first correlation index to obtain the bin-specific attention index corresponding to the terminal energy consumption sequence characteristics includes: Based on the terminal energy consumption sequence characteristics, determine the energy consumption characteristics and time characteristics corresponding to the terminal energy consumption sequence characteristics; Based on the energy consumption characteristics, the time characteristics, and the first correlation index, the bin attention index corresponding to the terminal energy consumption sequence characteristics is determined.

4. The method according to claim 3, characterized in that, The step of determining the bin attention feature corresponding to the terminal energy consumption sequence feature based on the terminal energy consumption sequence feature and the bin attention index includes: Based on the bin-specific attention index, a weighted summation operation is performed on the terminal energy consumption sequence features to obtain the bin-specific attention features corresponding to the terminal energy consumption sequence features.

5. The method according to claim 1, characterized in that, The step of determining the time-related information corresponding to the terminal energy consumption sequence characteristics based on the first time period and the second time period includes: Determine the time offset parameter between the first time period and the second time period; A time-related encoding operation is performed on the time offset parameter to obtain the time-related information corresponding to the terminal energy consumption sequence characteristics.

6. The method according to claim 1, characterized in that, The step of performing terminal energy consumption prediction based on at least two of the first features of interest to obtain terminal energy consumption prediction information includes: Perform multiple rounds of attention feature separation operation based on at least two first attention features, and obtain at least two second attention features obtained when performing the attention feature separation operation in each round; Perform a feature combination operation on the first feature of interest and all the second features of interest to obtain the energy-useful joint features; Based on the aforementioned joint energy consumption characteristics, terminal energy consumption prediction is performed to obtain terminal energy consumption prediction information.

7. The method according to claim 6, characterized in that, The feature separation operation performed in each round includes: Obtain at least two target energy consumption features, wherein the target energy consumption feature is the first feature of interest, or the second feature of interest obtained when the feature of interest separation operation was performed in the previous round; For each of the target energy consumption characteristics, a second correlation index is obtained between the terminal energy consumption sequence that matches the target energy consumption characteristic and the terminal energy consumption sequence that matches the associated energy consumption characteristic of the target energy consumption characteristic. Based on the target energy consumption characteristic and the second correlation index, a second weight index corresponding to the target energy consumption characteristic is determined. Based on the second weighting index, at least two second features of interest are determined from all the target energy consumption features.

8. An electronic device, characterized in that, include: A storage device on which computer programs are stored; A processing device for executing the computer program in the storage device to implement the steps of the method according to any one of claims 1-7.

Images