Air conditioner, control method and device thereof, storage medium and computer program product

Abstract

The application provides an air conditioner and a control method, device, storage medium and computer program product thereof, the method comprising: acquiring a plurality of groups of user physiological state parameters and environmental parameters in a pre-acquired environment and corresponding air conditioner running scene modes to obtain an original data set; preprocessing data in the acquired original data set to obtain N samples, each sample containing m features; using the preprocessed data to train a model, removing one feature at a time in each training, retaining the remaining m-1 features to train the model, and determining at least one feature with the smallest correlation with the prediction result, removing the at least one feature with the smallest correlation with the prediction result to obtain a prediction model; inputting the current user physiological state parameters and environmental parameters in the environment into the prediction model to predict the current air conditioner running scene mode. The application scheme can optimize the accuracy of the air conditioner algorithm model in predicting the scene.

Description

TECHNICAL FIELD

[0001] The present application relates to the control field, and particularly to an air conditioner and a control method, device, storage medium and computer program product thereof. BACKGROUND

[0002] With the wide application of intelligent air conditioning systems in modern buildings, the performance of system control models directly affects energy efficiency and user comfort. However, the intelligent air conditioning control models in the related art generally use a fixed parameter set or a strategy based on artificial experience. This static parameter configuration method is difficult to adapt to the dynamic changes in the actual operating environment. In actual application, the air conditioning system will be affected by various factors such as indoor and outdoor temperature changes, user behavior patterns, and equipment aging. These factors will cause dynamic changes in system parameters. Therefore, the control model lacking dynamic parameter optimization capability often leads to a decline in system performance, making it difficult to achieve precise control.

[0003] From the perspective of technical implementation, the method in the related art has obvious limitations in parameter optimization. Some studies attempt to perform correlation analysis through statistical methods to screen key parameters, but this method often ignores the time series characteristics and dynamic interaction between parameters. For example, in an air conditioning system, there are complex time sequence correlations between multiple parameters such as temperature, humidity, and wind speed. Parameter analysis at a single time point cannot accurately reflect the system dynamics. This parameter screening method based on static correlation analysis is prone to cumulative model errors, which in turn affects the control accuracy of the system. SUMMARY

[0004] The main purpose of the present application is to overcome the defects of the above-mentioned related art, and to provide an air conditioner and a control method, device, storage medium and computer program product thereof, to solve the problem that the air conditioning control model in the related art has obvious limitations in parameter optimization and dynamic adjustment, and is difficult to achieve precise control.

[0005] On the one hand, the present invention provides a method for controlling an air conditioner, comprising: obtaining a plurality of pre-collected sets of user physiological state parameters and environmental parameters in the environment and corresponding air conditioner operation scene modes to obtain an original data set; pre-processing the data in the obtained original data set to obtain N samples, each sample containing m features, wherein each set of user physiological state parameters and environmental parameters and corresponding air conditioner operation scene modes in the original data set is used as one sample, and the user physiological state parameters and environmental parameters in each sample are used as features; using the pre-processed data to perform model training to obtain a prediction model for predicting the air conditioner operation scene mode; wherein, each training step eliminates one feature of the m features in turn, retains the remaining m-1 features for model training, and evaluates and determines at least one feature with the smallest correlation with the prediction result, removes at least one feature with the smallest correlation with the prediction result, and obtains the prediction model; collecting the current user physiological state parameters and environmental parameters in the environment and inputting them into the prediction model to predict the current air conditioner operation scene mode, so as to control the operation of the air conditioner according to the predicted current air conditioner operation scene mode.

[0006] Optionally, each time training is performed, one of the m features is eliminated in turn, and the remaining m-1 features are retained for model training, and at least one feature with the smallest correlation with the prediction result is evaluated and determined, including: each time training is performed, one of the m features is eliminated in turn, and the remaining m-1 features are retained, and the data remaining after eliminating one feature is divided into a training set and a test set according to a preset ratio, and the training set is input into the prediction model for training, and the prediction model after training generates a prediction result through the test set, and calculates the evaluation index A according to the prediction result until all features have been eliminated; and the evaluation index A calculated according to the prediction result after each training is completed is compared, and at least one feature with the largest evaluation index after elimination is determined as the at least one feature with the smallest correlation with the prediction result.

[0007] Optionally, calculating the evaluation index according to the prediction result includes: calculating the evaluation index A by the following formula:

[0008] ;

[0009] Where N is the number of samples and Nr is the number of correctly predicted samples.

[0010] Optionally, further comprising: acquiring the user physiological state parameters and the environment parameters in the environment at every preset time within a preset time, and the air conditioner operation scene mode predicted by the prediction model; retraining the model according to the acquired user physiological state parameters and the environment parameters in the environment at every preset time within the preset time, and the air conditioner operation scene mode predicted by the prediction model, to update the input features of the model.

[0011] Optionally, the acquired data in the original data set is preprocessed, including:

[0012] The data in the original data set is standardized, and / or the data in the original data set is subjected to outlier rejection and / or missing value filling.

[0013] Optionally, the user physiological state parameters include at least one of body temperature, heart rate, blood pressure, respiratory rate, and exercise amount; and / or the environment parameters include at least one of temperature, humidity, air quality, wind speed, wind direction, noise, PM2.5 concentration, carbon dioxide concentration, carbon monoxide concentration, air conditioner operation mode, and air flow distribution.

[0014] Another aspect of the present application provides a control device of an air conditioner, comprising: an acquisition unit configured to acquire a plurality of groups of user physiological state parameters and environment parameters in an environment and corresponding air conditioner operation scene modes pre-acquired, to obtain an original data set; a processing unit configured to preprocess data in the original data set acquired by the acquisition unit, to obtain N samples, each sample containing m features, wherein each group of user physiological state parameters and environment parameters and the corresponding air conditioner operation scene mode in the original data set is taken as one sample, and the user physiological state parameters and the environment parameters in each sample are taken as features; a training unit configured to train a model using the preprocessed data of the processing unit, to obtain a prediction model for predicting an air conditioner operation scene mode; wherein each time, one feature of the m features is sequentially removed, the remaining m-1 features are retained for model training, and at least one feature with the smallest correlation with the prediction result is determined and removed, to obtain the prediction model; a prediction unit configured to input current user physiological state parameters and environment parameters in an environment into the prediction model, to predict a current air conditioner operation scene mode, and to control operation of the air conditioner according to the predicted current air conditioner operation scene mode.

[0015] Optionally, the training unit, in each training, sequentially eliminates one of the m features, retains the remaining m-1 features for model training, and determines at least one feature with the smallest correlation with the prediction result, comprising: in each training, sequentially eliminating one of the m features, retaining the remaining m-1 features, dividing the data remaining after eliminating one feature into a training set and a test set according to a preset ratio, inputting the training set into the prediction model for training, and generating a prediction result by the trained prediction model through the test set, and calculating an evaluation index A according to the prediction result, until all features have been eliminated; comparing the evaluation index A calculated according to the prediction result after each training, determining at least one feature with the largest evaluation index after elimination as at least one feature with the smallest correlation with the prediction result.

[0016] Optionally, the training unit calculates the evaluation index according to the prediction result, comprising: calculating the evaluation index A by the following formula:

[0017] ;

[0018] Wherein, N is the sample quantity, and Nr is the correct prediction sample quantity.

[0019] Optionally, the acquisition unit is further configured to: acquire the user physiological state parameters and the environment parameters in the environment collected in a preset time and the air conditioner running scene mode predicted by the prediction model every preset time; and the training unit is further configured to: re-perform model training according to the user physiological state parameters and the environment parameters in the environment collected in the preset time and the air conditioner running scene mode predicted by the prediction model acquired by the acquisition unit, to update the input features of the model.

[0020] Optionally, the processing unit pre-processes the data in the acquired original data set, comprising: standardizing the data in the original data set; and / or, eliminating outliers and / or filling missing values of the data in the original data set.

[0021] Optionally, the user physiological state parameters comprise at least one of body temperature, heart rate, blood pressure, respiratory rate, and exercise amount; and / or, the environment parameters comprise at least one of temperature, humidity, air quality, wind speed, wind direction, noise, PM2.5 concentration, carbon dioxide concentration, carbon monoxide concentration, air conditioner running mode, and air flow distribution.

[0022] In still another aspect, the present application provides a storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the steps of any of the above-mentioned methods.

[0023] In still another aspect, the present application provides an air conditioner comprising a processor, a memory, and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of any of the above-mentioned methods when executing the program.

[0024] In still another aspect, the present application provides an air conditioner comprising the above-mentioned control device.

[0025] In still another aspect, the present application provides a computer program product comprising a computer program, wherein the computer program implements the steps of any of the above-mentioned methods when executed by a processor.

[0026] According to the technical solution of the present application, by constructing an m*n parameter-time matrix as a reference input, based on the correlation quantification mechanism of the single-variable elimination method, a new input matrix is generated by eliminating one feature variable each time, and m times of parallel prediction is performed by using a neural network algorithm model, the variable with the lowest correlation is located by comparing the prediction results, and the feature variable with the smallest correlation is eliminated, so as to optimize the scene accuracy of model prediction and reduce the error of real-time regulation and control of the air conditioner scene mode. BRIEF DESCRIPTION OF DRAWINGS

[0027] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0028] Figure 1 is a method schematic diagram of an embodiment of the control method of the air conditioner provided by the present application;

[0029] Figure 2 is a method schematic diagram of another embodiment of the control method of the air conditioner provided by the present application;

[0030] Figure 3 shows a running logic schematic diagram of a specific embodiment of the air conditioner control method provided by the present application;

[0031] Figure 4 is a flowchart of a specific embodiment of the control method of the air conditioner provided by the present application;

[0032] Figure 5 is a structural block diagram of an embodiment of the control device of the air conditioner provided by the present application. DETAILED DESCRIPTION

[0033] In order to make the objects, technical solutions and advantages of the present application clearer, the following will clearly and completely describe the technical solutions of the present application with specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work belong to the protection scope of the present application.

[0034] It should be noted that the terms "first", "second" and the like in the description and claims of the present application and the above drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily have to include only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0035] From the perspective of technical implementation, the method in the related art has obvious limitations in parameter optimization. Some research attempts to screen key parameters through statistical methods for correlation analysis, but this method often ignores the time series characteristics and dynamic interaction between parameters. For example, in an air conditioning system, there are complex time series correlations between temperature, humidity, wind speed and other parameters, and parameter analysis at a single time point cannot accurately reflect the system dynamics. This parameter screening method based on static correlation analysis is prone to cumulative model errors, which further affects the control accuracy of the system.

[0036] In addition, the control model in the related art shows obvious low computational efficiency when facing the problem of parameter redundancy. In practical applications, an air conditioning system usually involves dozens or even hundreds of parameters, and these parameters may have high correlation. The model in the related art often needs to analyze and calculate all parameters one by one, resulting in waste of computing resources and increase of processing delay. At the same time, due to the lack of quantitative analysis of the real-time contribution of each parameter to the prediction result, the system is difficult to achieve accurate dynamic control, and is prone to control response lag and energy consumption increase. This lag effect not only affects the user's comfort experience, but also leads to energy waste, which is contrary to the original intention of energy saving and consumption reduction of the intelligent air conditioning system.

[0037] In conclusion, the intelligent air conditioner regulation model in the related art has obvious limitations in parameter optimization and dynamic adjustment, and a new regulation method capable of capturing dynamic changes in parameters in real time and quantifying the contribution of each parameter is urgently needed to improve the regulation accuracy and response speed of the system and realize truly intelligent and energy-saving operation.

[0038] The application provides a control method of an air conditioner.

[0039] Figure 1 The application provides a control method of an air conditioner.

[0040] As Figure 1 According to one embodiment of the application, the control method of the air conditioner at least includes steps S110, S120, S130 and S140.

[0041] In step S110, a plurality of sets of user physiological state parameters and environmental parameters in the environment and corresponding air conditioner operation scene modes are acquired in advance to obtain an original data set.

[0042] The user physiological state parameters can specifically include at least one of body temperature, heart rate, blood pressure, respiratory rate and exercise amount; and the environmental parameters can specifically include at least one of temperature, humidity, air quality, wind speed, wind direction, noise, PM2.5 concentration, carbon dioxide concentration, carbon monoxide concentration, air conditioner operation mode and air distribution. The user physiological state parameters can be collected by wearable devices such as smart bracelets and smart watches. The environmental parameters can be detected by sensors installed on the air conditioner, such as temperature sensors and humidity sensors. The user physiological state parameters and environmental parameters in each set are collected at the same time.

[0043] The corresponding air conditioner operation scene mode is determined according to the user physiological state parameters and is used as a label y. For example, the air conditioner operation scene mode corresponding to the user physiological state parameters can be manually labeled in advance. The air conditioner operation scene mode can specifically include at least one of a sports scene, a pathological state, a sleep scene and an allergy scene.

[0044] For example, in combination with wearable devices and built-in health modules of air conditioners, user end data (body temperature, heart rate, blood pressure, respiratory rate, exercise amount, etc.) is collected by wearable devices such as smart bracelets and smart watches, sensor modules are installed on air conditioner units to monitor environmental parameters (such as temperature, humidity, noise, wind speed, etc.), the health status of the user (such as normal, sports, rest, etc.) is preliminarily set according to the collected data, and labeling is completed.

[0045] Step S120, the data in the original data set obtained is preprocessed to obtain N samples, each sample containing m features.

[0046] Specifically, each set of user physiological state parameters and environmental parameters and the corresponding air conditioner running scene mode in the original data set is taken as one sample, the user physiological state parameters and environmental parameters in each sample are taken as features, i.e., as inputs of the model, and the corresponding air conditioner running scene mode is taken as the output of the model. The original data set D is preprocessed, for example, the data set D contains N samples, each sample contains m features. The features include user physiological state parameter features and environmental parameter features, the user physiological state parameter features can specifically include at least one of user body temperature x1, heart rate x2, blood pressure x3, respiratory rate x4, and exercise amount x5; the environmental parameter features can specifically include at least one of temperature x6, humidity x7, air quality x8, wind speed x9, wind direction x 10 , noise x 11 , PM2.5 concentration x 12 , carbon dioxide concentration x 13 , carbon monoxide concentration x 14 , air conditioner running mode x 15 , air flow distribution x 16 , and the air conditioner running scene mode (for example, including at least one of the exercise scene y1, the pathological state y2, the sleep scene y3, and the allergy scene y4) is taken as a classification label, denoted by label y, i.e., the output of the model.

[0047] In order to make the features x i on the same scale, the data in the original data set is standardized. Assuming that the mean of the feature value x i is μ i , and the standard deviation is σ i , the standardized feature x i ' can be expressed as:

[0048] ;

[0049] Optionally, before standardization, the data in the original data set can also be subjected to outlier rejection and / or missing value filling. For example, outliers that deviate significantly from the normal range are deleted, or statistical quantities such as mean, median, etc. are used to replace outliers. For example, the mean, median or mode can be used to fill in the missing values, keeping the data distribution characteristics, or the missing values can be filled in by linear interpolation or polynomial interpolation.

[0050] Step S130, using the preprocessed data to train a model to obtain a prediction model for predicting the air conditioner running scene mode.

[0051] Specifically, the pre-processed data contains N samples, each of which contains m features. Each time, one feature of the m features is removed, and the remaining m-1 features are retained for model training, and the at least one feature with the smallest correlation with the prediction result is determined and removed, and the remaining features are retained as input features of the prediction model, to obtain the prediction model.

[0052] Preferably, the pre-processed data is divided into a training set C and a test set T according to a preset ratio, for example, the training set C and the test set T are divided according to a ratio of 7:3. The training set C is used as input to the prediction model for training, and the trained model is used to generate a prediction result in the test set T, and an evaluation index A is used to calculate and determine the result of the model:

[0053] ;

[0054] Wherein, N is the total number of samples, and Nr is the number of correctly predicted samples.

[0055] Specifically, each time, one feature of the m features is removed, and the remaining m-1 features are retained, and the remaining data after removing one feature is divided into a training set and a test set according to a preset ratio, and the training set is input into the prediction model for training, and the trained prediction model generates a prediction result through the test set, and the evaluation index A is calculated according to the prediction result, until all the features have been removed, and m times of training are completed. By comparing the evaluation indexes A calculated according to the prediction results after each training, at least one feature with the maximum evaluation index after removal is determined as at least one feature with the smallest correlation with the prediction result. Since the evaluation index represents the accuracy of the prediction model, if the accuracy of the prediction model is higher after removing a certain feature, it means that the correlation between the feature and the prediction result is smaller, so at least one feature with the smallest correlation with the prediction result can be removed from the input features of the model, that is, at least one feature with the maximum evaluation index after removal.

[0056] Step S140, collect the current user physiological state parameters and environmental parameters in the environment and input them into the prediction model to predict the current air conditioner running scene mode, so as to control the operation of the air conditioner according to the predicted current air conditioner running scene mode.

[0057] Specifically, according to the remaining features after removing at least one feature with the smallest correlation with the prediction result, the current user physiological state parameters and environmental parameters in the environment are collected and input into the prediction model to predict the current air conditioner running scene mode, so as to control the operation of the air conditioner according to the current air conditioner running scene mode.

[0058] According to the air conditioner running scene mode, a corresponding control strategy can be used to control the running of the air conditioner. For example, when the air conditioner running scene mode is the sports mode, the rapid cooling (such as increasing the compressor running frequency and / or increasing the inner fan rotating speed) can be started, and the fresh air system can be linked to perform fresh air replacement; when the air conditioner running scene mode is the pathological mode, the cabinet ultraviolet sterilization function can be started, and the humidity can be adjusted to the preset humidity range (for example, the humidity is adjusted by the humidifying function or the dehumidifying function of the air conditioner); when the air conditioner running scene mode is the night mode, the quiet mode is adjusted to avoid interrupting the user's sleep.

[0059] Figure 2 is a method schematic diagram of another embodiment of the air conditioner control method provided by the application.

[0060] As shown in Figure 2 According to another embodiment of the application, the air conditioner control method further includes steps S150 and S160.

[0061] In step S150, the user physiological state parameters and the environment parameters in the environment where the user is located collected in the preset time and the air conditioner running scene mode predicted by the prediction model are obtained every preset time.

[0062] In step S160, the user physiological state parameters and the environment parameters in the environment where the user is located collected in the preset time and the air conditioner running scene mode predicted by the prediction model are obtained every preset time.

[0063] Specifically, the feature screening is triggered every preset time, the user physiological state parameters and the environment parameters in the environment where the user is located collected in the preset time and the air conditioner running scene mode predicted by the prediction model are obtained, and the model training is performed again, at least one feature with the smallest correlation with the prediction result is removed, and the input features of the prediction model are updated. For example, the feature screening is triggered every 6 hours, the user physiological state parameters and the environment parameters collected in the previous 6 hours and the air conditioner running scene mode predicted by the prediction model are obtained, and the model training is performed again.

[0064] To clearly illustrate the technical scheme of the application, the execution flow of the air conditioner control method provided by the application is described below with one specific embodiment.

[0065] Figure 3 The running logic schematic diagram of one specific embodiment of the air conditioner control method provided by the application is shown. As shown in Figure 3As shown, the data acquisition part collects user physiological state parameters through the wearable device, including body temperature, heart rate, blood pressure, respiratory rate, exercise amount, etc., collects indoor related parameters through the machine group sensor, and determines the corresponding air conditioner running scene mode. The individual mode establishment part uses the collected data to train the prediction model, divides the preprocessed data into a training set and a test set for model training; the intelligent response part collects user physiological state parameters and environmental parameters in real time, judges the current air conditioner running scene mode through the prediction model, and controls the operation of the air conditioner according to the predicted air conditioner running scene mode.

[0066] Figure 4 is a flowchart of a specific embodiment of the control method of the air conditioner provided by the application. As shown in Figure 4 When the air conditioner is running, the user physiological state parameters and the indoor related parameters (environmental parameters) are collected, the collected data is preprocessed to extract the characteristic value Xi, the data preprocessing includes abnormal data value judgment and elimination, missing value filling and data normalization, a characteristic value Xi is sequentially eliminated, training and testing are performed, the evaluation index A is output, each evaluation index A is sequentially compared, the correlation of each characteristic value is output, the characteristic with the smallest correlation is eliminated, and the input characteristic of the prediction model is dynamically updated every 6 hours. Real-time collection of user physiological state parameters and indoor related parameters (environmental parameters), judgment of the current air conditioner running scene mode through the prediction model, and control of the operation of the air conditioner according to the predicted current air conditioner running scene mode.

[0067] The application also provides a control device of an air conditioner.

[0068] Figure 5 is a structural block diagram of an embodiment of the control device of the air conditioner provided by the application. As shown in Figure 5 The control device 100 of the air conditioner comprises an acquisition unit 110, a processing unit 120, a training unit 130 and a prediction unit 140.

[0069] The acquisition unit 110 is used for acquiring a plurality of groups of user physiological state parameters and environmental parameters in the environment and corresponding air conditioner running scene modes collected in advance, to obtain an original data set.

[0070] The user physiological state parameters can specifically include at least one of body temperature, heart rate, blood pressure, respiratory rate, and exercise amount; and the environment parameters can specifically include at least one of temperature, humidity, air quality, wind speed, wind direction, noise, PM2.5 concentration, carbon dioxide concentration, carbon monoxide concentration, air conditioner operation mode, and air flow distribution. The user physiological state parameters can be collected by a wearable device such as a smart bracelet or a smart watch. The environment parameters can be detected by sensors installed on the air conditioner, such as temperature sensors and humidity sensors. The user physiological state parameters and the environment parameters in each of the multiple groups of environments are collected at the same time. The user physiological state parameters and the environment parameters in each of the multiple groups of environments are collected at the same time.

[0071] The corresponding air conditioner operation scene mode is determined according to the user physiological state parameters, and is taken as a label y. For example, the air conditioner operation scene mode corresponding to the user physiological state parameters can be pre-labeled by manual labeling. The air conditioner operation scene mode can specifically include at least one of an exercise scene, a pathological state, a sleep scene, and an allergy scene.

[0072] For example, in combination with a wearable device and a health module built in an air conditioner, user end data (body temperature, heart rate, blood pressure, respiratory rate, exercise amount, etc.) is collected by the wearable device such as a smart bracelet or a smart watch, environment parameters (such as temperature, humidity, noise, wind speed, etc.) are monitored by installing a sensor module on the air conditioner unit, and the health status of the user (such as normal, exercise, rest, etc.) is preliminarily set according to the collected data, and labeling is completed.

[0073] The processing unit 120 is configured to pre-process data in the original data set obtained by the acquisition unit 110 to obtain N samples, each sample containing m features.

[0074] Specifically, each group of user physiological state parameters and environment parameters and the corresponding air conditioner operation scene mode in the original data set are taken as one sample, the user physiological state parameters and the environment parameters in each sample are taken as features, that is, as the input of the model, and the corresponding air conditioner operation scene mode is taken as the output of the model. The original data set D is pre-processed, for example, the data set D contains N samples, and each sample contains m features. The features include user physiological state parameter features and environment parameter features. The user physiological state parameter features can specifically include at least one of user body temperature x1, heart rate x2, blood pressure x3, respiratory rate x4, and exercise amount x5; and the environment parameter features can specifically include temperature x6, humidity x7, air quality x8, wind speed x9, wind direction x 10 , noise x 11 , PM2.5 concentration x12 carbon dioxide concentration x 13 carbon monoxide concentration x 14 air conditioning operation mode x 15 airflow distribution x 16 at least one of the air conditioning operation scene mode (for example, including at least one of the sports scene y1, the pathological state y2, the sleep scene y3, and the allergy scene y4) is taken as a classification label, denoted as label y, that is, the model output.

[0075] In order to make the features x i The data in the original data set is standardized on the same scale. Assuming that the mean of the feature value x i is μ i , and the standard deviation is σ i , the standardized feature x i ' can be expressed as:

[0076] ;

[0077] Optionally, before standardization, the data in the original data set can also be subjected to outlier rejection and / or missing value filling. For example, outliers that deviate significantly from the normal range are deleted, or statistical quantities such as mean, median, etc. are used to replace outliers. For example, the mean, median or mode can be used to fill in the missing values, maintain the data distribution characteristics, or the missing values can be filled in by linear interpolation or polynomial interpolation.

[0078] The training unit 130 is configured to train a model using the preprocessed data from the processing unit 120 to obtain a prediction model for predicting the air conditioning operation scene mode.

[0079] Specifically, the preprocessed data contains N samples, each containing m features. Each time, one feature of the m features is removed, and the remaining m-1 features are retained for model training, and the correlation between the prediction result is determined and at least one feature with the smallest correlation is removed. The remaining features are retained as input features of the prediction model to obtain the prediction model.

[0080] Preferably, the preprocessed data is divided into a training set C and a test set T according to a preset ratio, for example, the training set C and the test set T are divided according to a ratio of 7:3. The training set C is used as input to the prediction model for training, and the trained model is used to generate a prediction result in the test set T, and an evaluation index A is used to calculate and determine the result of the model:

[0081] ;

[0082] Wherein, N is the total number of samples, and Nr is the number of correctly predicted samples.

[0083] Specifically, each time, one of the m features is removed, and the remaining m-1 features are retained. The data remaining after removing one feature is divided into a training set and a test set according to a preset ratio. The training set is input into the prediction model for training. The prediction model after training generates a prediction result through the test set. An evaluation index A is calculated according to the prediction result, until all features have been removed, and m times of training are completed. The evaluation index A calculated according to the prediction result after each time of training is compared, and at least one feature with the maximum evaluation index after removal is determined as at least one feature with the minimum correlation with the prediction result. Since the evaluation index represents the accuracy of the prediction model, if the accuracy of the prediction model is higher after removing a certain feature, it indicates that the correlation of the feature with the prediction result is smaller, and therefore at least one feature with the minimum correlation with the prediction result, i.e., at least one feature with the maximum evaluation index after removal, can be removed from the model input features.

[0084] The prediction unit 140 is configured to collect current user physiological state parameters and environmental parameters in the environment and input the prediction model to predict a current air conditioner running scene mode, so as to control the operation of the air conditioner according to the predicted current air conditioner running scene mode.

[0085] Specifically, according to the features remaining after removing at least one feature with the minimum correlation with the prediction result, current user physiological state parameters and environmental parameters in the environment are collected and input into the prediction model to predict a current air conditioner running scene mode, so as to control the operation of the air conditioner according to the current air conditioner running scene mode.

[0086] The air conditioner can be controlled according to the corresponding control strategy according to the air conditioner running scene mode. For example, when the air conditioner running scene mode is the sports mode, the rapid cooling (such as increasing the compressor operation frequency and / or increasing the inner fan speed) can be started, and the fresh air system can be linked to perform fresh air replacement; when the air conditioner running scene mode is the pathological mode, the cabinet ultraviolet sterilization function can be started, and the humidity can be adjusted to a preset humidity range (for example, the humidity is adjusted by the humidifying function or the dehumidifying function of the air conditioner); when the air conditioner running scene mode is the night mode, the air conditioner is adjusted to the silent mode to avoid interrupting the user's sleep.

[0087] Optionally, the acquisition unit 110 is further configured to acquire the user physiological state parameters and the environmental parameters in the environment at a preset time interval, and the air conditioner operation scenario mode predicted by the prediction model; and the training unit 130 is further configured to re-train the model according to the user physiological state parameters and the environmental parameters in the environment at the preset time interval, and the air conditioner operation scenario mode predicted by the prediction model, so as to update the input features of the model.

[0088] Specifically, the feature screening is triggered at a preset time interval, the user physiological state parameters and the environmental parameters in the environment at the time interval are acquired, and the air conditioner operation scenario mode predicted by the prediction model is acquired, the model is re-trained, at least one feature with the minimum correlation with the prediction result is removed, and the input features of the prediction model are updated. For example, the feature screening is triggered every 6 hours, the user physiological state parameters and the environmental parameters in the previous 6 hours are acquired, the air conditioner operation scenario mode predicted by the prediction model is acquired, the model is re-trained.

[0089] The application further provides a storage medium corresponding to the control method of the air conditioner, and the storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of any one of the foregoing methods.

[0090] The application further provides an air conditioner corresponding to the control method of the air conditioner, and the air conditioner comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of any one of the foregoing methods when executing the computer program.

[0091] The application further provides an air conditioner corresponding to the control device of the air conditioner, and the air conditioner comprises the control device of any one of the foregoing air conditioners.

[0092] The application further provides a computer program product corresponding to the control method of the air conditioner, and the computer program product comprises a computer program, and the computer program is executed by a processor to implement the steps of any one of the foregoing methods.

[0093] According to the scheme, the m*n parameter-time matrix is constructed as a reference input, the correlation quantification mechanism based on the single-variable elimination method is used, one feature variable is eliminated each time to generate a new input matrix, the neural network algorithm model is used for m times of parallel prediction, the variables with the lowest correlation are located by comparing the prediction results, and the feature variable with the minimum correlation is eliminated, so that the model prediction scene accuracy is optimized, and the error of real-time regulation of the air conditioner scene mode is reduced.

[0094] The air conditioning system in the related art cannot perceive the correlation between the user physiological state and the environment, and the algorithm model commonly used in the fixed operation mode has low prediction scene accuracy, is prone to error control, and often causes environmental regulation response delay and body discomfort. The air conditioning system with adaptive learning ability is realized by fusing environmental data and user physiological state parameters. Based on variable correlation analysis, the accuracy of the air conditioning algorithm model prediction scene is optimized, error control is reduced, and the user experience and the intelligent degree of the air conditioning are effectively improved.

[0095] The present application scheme can be extended to the air conditioning system of smart home, commercial building, etc., to improve the environmental regulation precision and energy utilization rate. Combined with the Internet of Things platform, accurate temperature and humidity control of special scenes such as hospitals and laboratories can be realized to ensure environmental safety. For the production of variable frequency air conditioners, the technical scheme of the present application can be deployed through OTA upgrade, without the need to replace hardware to improve product competitiveness.

[0096] The functions described herein can be implemented in hardware, software executed by a processor, firmware, or any combination thereof. If implemented in software executed by a processor, the functions can be stored on or transferred over as one or more instructions or code on a computer-readable medium. Other examples and implementations are within the scope and spirit of the disclosure and appended claims. For example, due to the nature of software, functions described above can be implemented using software executed by a processor, hardware, firmware, hardwiring, or combinations of any of these. Also, each of the functions can be implemented as a separate function or combined as a single function in a process unit. Further, the functions can be implemented in a single device or distributed among several devices.

[0097] In several embodiments provided in the present application, it should be understood that the disclosed technical content can be implemented by other ways. Among them, the above-mentioned device embodiments are only schematic, for example, the division of the units can be a logical function division, and actual implementation can have another division way, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the coupling or direct coupling or communication connection between the units or modules shown or discussed can be indirect coupling or communication connection through some interfaces, units or modules, which can be electrical or other forms.

[0098] The units described as separate components can or can not be physically separated, and the components of the control device can or can not be physical units, i.e. can be located in one place or distributed on multiple units. Part or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment scheme.

[0099] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the entire or part of the technical solutions that make contributions to the related art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0100] The above merely describes the embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1. A method for controlling an air conditioner, characterized in that: include: Acquire multiple sets of pre-collected user physiological state parameters and environmental parameters in the environment and corresponding air-conditioning operation scene modes to obtain an original data set; Preprocessing the data in the obtained original data set to obtain N samples, each sample containing m features, wherein each set of user physiological state parameters and environmental parameters and the corresponding air conditioning operation scene mode in the original data set is used as one sample, and the user physiological state parameters and environmental parameters in each sample are used as features; Performing model training using the preprocessed data to obtain a prediction model for predicting the air conditioner operation scene mode; wherein, during each training, one of the m features is sequentially removed, and the remaining m-1 features are retained for model training, and at least one feature having the smallest correlation with the prediction result is evaluated and determined, and the at least one feature having the smallest correlation with the prediction result is removed to obtain the prediction model; Collecting current user physiological state parameters and environmental parameters in the environment and inputting them into the prediction model to predict the current air-conditioning operation scene mode, and controlling the operation of the air-conditioning according to the predicted current air-conditioning operation scene mode; Each time the training is performed, one of the m features is removed in turn, and the remaining m-1 features are retained for model training. At least one feature with the least correlation with the prediction result is evaluated and determined, including: Each time training, one of the m features is removed in turn, and the remaining m-1 features are retained. The remaining data after removing one feature is divided into a training set and a test set according to a preset ratio. The training set is input into the prediction model for training. After the training, the prediction model generates a prediction result through the test set, and the evaluation index A is calculated based on the prediction result, until all features have been removed; The evaluation index A calculated according to the prediction result after each training is completed is compared, and at least one feature with the largest evaluation index after elimination is determined as the at least one feature with the smallest correlation with the prediction result.

2. The method according to claim 1, characterized in that Calculate the evaluation index based on the prediction results, including: calculating the evaluation index A by the following formula: A=Nr / N; Where N is the number of samples and Nr is the number of correctly predicted samples.

3. The method according to claim 1 or 2, characterized in that Also includes: Acquire, at preset intervals, user physiological state parameters and environmental parameters collected within the preset time period, and the air conditioning operation scene mode predicted by the prediction model; The model is retrained based on the user's physiological state parameters and environmental parameters collected in the preset time and the air-conditioning operation scene mode predicted by the prediction model.

4. The method according to claim 1 or 2, characterized in that Preprocessing the data in the collected original data set includes: Standardizing the data in the original data set; and / or, Outliers are removed and / or missing values ​​are filled in for the data in the original data set.

5. The method according to claim 1 or 2, characterized in that The user's physiological state parameter includes at least one of body temperature, heart rate, blood pressure, respiratory rate, and exercise volume; and / or, The environmental parameters include: at least one of temperature, humidity, air quality, wind speed, wind direction, noise, PM2.5 concentration, carbon dioxide concentration, carbon monoxide concentration, air conditioning operation mode, and air flow distribution.

6. A control device for an air conditioner, characterized in that: include: An acquisition unit is used to acquire a plurality of pre-collected sets of user physiological state parameters and environmental parameters in the environment and corresponding air-conditioning operation scene modes to obtain an original data set; a processing unit, configured to preprocess the data in the original data set acquired by the acquisition unit to obtain N samples, each sample comprising m features, wherein each set of user physiological state parameters and environmental parameters and the corresponding air conditioning operation scene mode in the original data set is regarded as one sample, and the user physiological state parameters and environmental parameters in each sample are regarded as features; a training unit, configured to perform model training using the data preprocessed by the processing unit to obtain a prediction model for predicting the air conditioner operation scene mode; wherein, each training step sequentially removes one of the m features, retains the remaining m-1 features for model training, evaluates and determines at least one feature having the least correlation with the prediction result, and removes the at least one feature having the least correlation with the prediction result to obtain the prediction model; a prediction unit, configured to collect current user physiological state parameters and environmental parameters in the environment and input them into the prediction model to predict a current air-conditioning operation scene mode, so as to control the operation of the air-conditioning according to the predicted current air-conditioning operation scene mode; The training unit sequentially removes one feature from the m features during each training session, retains the remaining m-1 features for model training, and evaluates and determines at least one feature with the smallest correlation with the prediction result, including: sequentially removing one feature from the m features during each training session, retaining the remaining m-1 features, dividing the remaining data after removing one feature into a training set and a test set according to a preset ratio, inputting the training set into the prediction model for training, generating a prediction result through the test set for the trained prediction model, and calculating an evaluation index A based on the prediction result until all features have been removed; and comparing the evaluation index A calculated based on the prediction result after each training session, determining at least one feature with the largest evaluation index after removal as the at least one feature with the smallest correlation with the prediction result.

7. A storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

8. An air conditioner, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method according to any one of claims 1 to 5 are implemented, and the method comprises the control device according to claim 6.

9. A computer program product, characterized in that The invention comprises a computer program, which implements the steps of the method according to any one of claims 1 to 5 when the computer program is executed by a processor.

Images