Water purifier control system based on Internet of Things

Abstract

The invention relates to the technical field of water purifier control, and particularly discloses a water purifier control system based on Internet of Things, which is characterized in that a data acquisition module is used for acquiring a water use record of a target user and an Internet of Things monitoring record, then a first prediction module is used for acquiring a predicted total water use amount, and a primary control module is used for acquiring a predicted water purifier flushing amount; then the predicted water consumption frequency is obtained through the second prediction module, and finally the washing interval and the washing duration of the water purifier are adjusted through the secondary control module. Compared with the prior art, on one hand, the water consumption record and the internet-of-things monitoring record are used for predicting the total water consumption to achieve adjustment of the flushing amount, and on the other hand, the water consumption record and the internet-of-things monitoring record are used again for predicting the water consumption frequency to adjust the flushing interval and the flushing duration more finely. Therefore, the control strategy of the water purifier is flexibly changed according to the work and rest habits of the user, so that the water purifier can adapt to the use habits of the user, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of water purifier control, and in particular to a water purifier control system based on the Internet of Things. Background Art

[0002] A water purifier is a common water treatment device in modern homes. Its primary function is to remove impurities, bacteria, heavy metals, and other harmful substances from water through its built-in filtration system, thereby providing clean, safe drinking water. As people's pursuit of a healthier lifestyle continues to improve, water purifiers have become an indispensable part of many households.

[0003] The flushing function of a water purifier is a crucial maintenance mechanism, ensuring internal cleanliness and the long-lasting filtration effect. This function is typically performed either timed or manually. It works by backwashing the filter element with water to remove dirt and impurities adhering to the surface. This process not only extends the filter's lifespan but also maintains the purifier's filtration efficiency, ensuring consistent water quality.

[0004] However, in the prior art, the flushing control of the water purifier usually follows a preset fixed pattern, such as automatically flushing based on the flushing interval and duration set in advance by the manufacturer. Although this method simplifies the operating process, it sacrifices adaptability to actual usage conditions. In practice, due to the uncertainty of the usage frequency of different users, a fixed flushing cycle may not be able to effectively adapt to the work and rest habits of all users. This may not only reduce the cleaning efficiency of the filter element due to too few flushes, but also may cause unnecessary replacement of the filter element due to too many flushes. Therefore, people need a water purifier control solution that can flexibly adapt to the work and rest habits of users. Summary of the Invention

[0005] The purpose of this invention is to provide a water purifier control system based on the Internet of Things to solve the following technical problems: Existing water purifier control methods cannot flexibly adapt to user's work and rest habits.

[0006] The purpose of the present invention can be achieved through the following technical solutions: A water purifier control system based on the Internet of Things, comprising: Data acquisition module, used to obtain water consumption records and IoT monitoring records of target users; The first prediction module is used to predict the total water consumption of the target user within the target time period based on the water consumption record of the target user and the IoT monitoring record; The primary control module is used to obtain the predicted flushing volume of the water purifier by the target user within the target time period based on the predicted total water consumption; The second prediction module is used to predict the target user's water usage frequency within the target time period based on the target user's water usage record and IoT monitoring record; The secondary control module is used to adjust the flushing interval and flushing time of the water purifier according to the predicted water usage frequency and the predicted flushing volume of the water purifier, wherein the flushing interval is inversely proportional to the water usage frequency, and the flushing time is inversely proportional to the water usage frequency.

[0007] As a further solution of the present invention: the first prediction module includes: A first vector establishment module is configured to establish a first vector based on the target user's water consumption records and IoT monitoring records, wherein the first vector is used to represent the target user's water consumption characteristics and the environmental characteristics monitored by the IoT, wherein the element values ​​in the first vector are used to represent the historical total water consumption, the historical average water consumption change rate, and the various IoT monitoring records obtained based on the water consumption records; The first network calling module is used to input the first vector into the first preset neural network to obtain the predicted total water consumption output by the first preset neural network.

[0008] As a further solution of the present invention: the first preset neural network includes a first input layer, at least one first hidden layer and a first output layer connected in sequence, wherein the number of nodes in the first input layer is equal to the number of elements in the first vector, each node in the first input layer is used to input an element in the first vector, the number of first hidden layers is proportional to the number of elements in the first vector, and the number of nodes in the first output layer is one, which is used to output the predicted total water consumption.

[0009] As a further solution of the present invention: the number of the first hidden layer is determined by the following formula: ; Wherein, N is the number of first hidden layers, [] is a rounding symbol, e is a natural constant, x is the number of elements in the first vector, a is a preset scaling factor, and n is the preset number of nodes in each first hidden layer, where n is greater than or equal to x.

[0010] As a further solution of the present invention: the second prediction module includes: A second vector establishment module is used to establish a second vector sequence based on the target user's water usage record and the IoT monitoring record, wherein the second vector sequence includes multiple second vectors, each second vector corresponds to a time period, and the multiple second vectors are arranged based on the corresponding time sequence. The second vectors are used to represent the water usage characteristics of the target user in the corresponding time period and the environmental characteristics monitored by the IoT in the corresponding time period, wherein the element values ​​in the second vector are used to represent the corresponding time period, the number of water usage times in the time period obtained based on the water usage record, and the IoT monitoring record in the time period; The second network calling module is used to input the second vector sequence into the second preset neural network to obtain the predicted water use frequency output by the second preset neural network.

[0011] As a further solution of the present invention: the second preset neural network includes a second recurrent neural network and a third recurrent neural network, wherein the second recurrent neural network is connected to the third recurrent neural network, the second recurrent neural network is used to output associated hyperparameters according to the second vector sequence, the associated hyperparameters are used to describe the degree of correlation between the target user's water usage records and the Internet of Things monitoring records, and the third recurrent neural network is used to output the predicted water usage frequency according to the second vector sequence and the associated hyperparameters.

[0012] As a further solution of the present invention: the second recurrent neural network includes a second input layer, at least one second hidden layer and a second output layer connected in sequence, wherein the number of nodes in the second input layer is the same as the number of elements in the second vector, each node in the second input layer is used to input an element in the second vector, each node in the second input layer is connected to the node in the adjacent second hidden layer, and each node in the second hidden layer is also connected to other nodes in the same layer, the second hidden layer is used for cyclic prediction, the number of cycles of the second hidden layer is the same as the number of second vectors in the second vector sequence, the second output layer includes a node, the node of the second output layer is connected to all nodes in the connected second hidden layer, and the node of the second output layer is used to output associated hyperparameters in the last cycle.

[0013] As a further solution of the present invention: the third cyclic neural network includes a third input layer, a third hidden layer and a third output layer connected in sequence, wherein the number of nodes in the third input layer is the same as the number of elements in the second vector, each node in the third input layer is used to input an element of the second vector, each node in the third input layer is respectively connected to a node in the third hidden layer, and each node in the third hidden layer is also respectively connected to a node in the second output layer, the third hidden layer is used for cyclic prediction, the number of cycles of the third hidden layer is the same as the number of second vectors in the second vector sequence, the third output layer includes one node, the node of the third output layer is connected to all nodes in the third hidden layer, and the node of the third output layer is used to output the predicted water use frequency in the last cycle.

[0014] Beneficial effects of the present invention: The present invention provides a water purifier control system based on the Internet of Things. The control system first obtains the water consumption records and Internet of Things monitoring records of a target user through a data acquisition module, then predicts the total water consumption of the target user within a target time period based on the water consumption records and Internet of Things monitoring records of the target user through a first prediction module, then obtains the predicted water purifier flushing volume of the target user within the target time period based on the predicted total water consumption through a primary control module, then predicts the predicted water consumption frequency of the target user within the target time period based on the water consumption records and Internet of Things monitoring records of the target user through a second prediction module, and finally adjusts the flushing interval and flushing duration of the water purifier based on the predicted water consumption frequency and the predicted water purifier flushing volume according to the predicted water consumption frequency, wherein the flushing interval is inversely proportional to the water consumption frequency, and the flushing duration is inversely proportional to the water consumption frequency. Compared with the prior art, the present invention combines the Internet of Things technology, on the one hand, uses the water consumption records and Internet of Things monitoring records to predict the total water consumption, thereby realizing the adjustment of the flushing volume, and on the other hand, uses the water consumption records and Internet of Things monitoring records to predict the water consumption frequency again, so that when the flushing volume is determined, the flushing interval and flushing duration can be more finely adjusted, thereby realizing the flexible change of the control strategy of the water purifier according to the user's work and rest habits, so that it can adapt to the user's usage habits and improve the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 This is a schematic diagram of the system structure of the water purifier control system based on the Internet of Things of the present invention; Figure 2 It is a schematic diagram of the structure of the second preset neural network in the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 As shown, the present invention is a water purifier control system based on the Internet of Things, which is characterized by including: The data acquisition module 110 is used to obtain the water consumption records and IoT monitoring records of target users; The first prediction module 120 is used to predict the total water consumption of the target user within the target time period based on the water consumption record of the target user and the IoT monitoring record; The primary control module 130 is configured to obtain the predicted flushing volume of the water purifier by the target user within the target time period based on the predicted total water consumption; The second prediction module 140 is used to predict the target user's water usage frequency within the target time period based on the target user's water usage record and IoT monitoring record; The secondary control module 150 is used to adjust the flushing interval and flushing time of the water purifier according to the predicted water usage frequency and the predicted flushing volume of the water purifier, wherein the flushing interval is inversely proportional to the water usage frequency, and the flushing time is inversely proportional to the water usage frequency.

[0019] In the above context, water use records are data records of the target user's water use behavior over a period of time. These records can include water consumption, valve opening times, and any other record that can indicate water use. IoT monitoring records refer to IoT monitoring data related to the target user's environment and daily behavior, such as temperature, humidity, water quality, lighting activation records, smart home usage records, and door lock activation records. These IoT monitoring records can be obtained through smart homes or other pre-deployed sensing devices. The specific types of these IoT monitoring records can be flexibly determined based on the product positioning strategy and actual application of the water purifier in the actual application.

[0020] The target time period mentioned above is the future period of time during which water purifier control optimization is desired, such as the next week or the next month. The predicted water purifier flush volume is the total amount of flushes predicted for the water purifier during the target time period. It can be expressed in any dimension, such as the number of flushes, water consumption, or time.

[0021] Compared with the existing technology, the present invention combines the Internet of Things technology. On the one hand, it uses water consumption records and Internet of Things monitoring records to predict the total water consumption and realizes the adjustment of the flushing volume. On the other hand, it again uses water consumption records and Internet of Things monitoring records to predict the water consumption frequency. When the flushing volume is determined, the flushing interval and flushing time can be adjusted more finely, thereby realizing flexible changes in the control strategy of the water purifier according to the user's work and rest habits, so that it can adapt to the user's usage habits and improve the user experience.

[0022] Furthermore, in a preferred embodiment, the first prediction module 120 specifically includes: A first vector establishment module is configured to establish a first vector based on the target user's water consumption records and IoT monitoring records, wherein the first vector is used to represent the target user's water consumption characteristics and the environmental characteristics monitored by the IoT, wherein the element values ​​in the first vector are used to represent the historical total water consumption, the historical average water consumption change rate, and the various IoT monitoring records obtained based on the water consumption records; The first network calling module is used to input the first vector into the first preset neural network to obtain the predicted total water consumption output by the first preset neural network.

[0023] The above design represents the water usage characteristics of the target user in the past period and the environmental characteristics monitored by the Internet of Things in the form of a first vector, and comprehensively analyzes the above characteristics through a first preset neural network. The ability of neural networks to solve complex problems is used to obtain a more accurate prediction of the total water consumption.

[0024] Specifically, in a preferred embodiment, the first preset neural network includes a first input layer, at least one first hidden layer and a first output layer connected in sequence, wherein the number of nodes in the first input layer is equal to the number of elements in the first vector, each node in the first input layer is used to input an element in the first vector, the number of first hidden layers is proportional to the number of elements in the first vector, and the number of nodes in the first output layer is one, which is used to output the predicted total water consumption.

[0025] As mentioned above, the types of IoT monitoring records that can be collected in practice need to be determined according to the specific situation, and the difficulty of preparing sample data when training the first preset neural network also needs to be considered. Therefore, the embodiment does not impose specific restrictions on the number of first hidden layers, and the number of first hidden layers can be flexibly determined according to the specific situation.

[0026] Specifically, in a preferred embodiment, the number of the first hidden layer is determined by the following formula: ; Wherein, N is the number of first hidden layers, [] is a rounding symbol, e is a natural constant, x is the number of elements in the first vector, a is a preset scaling factor, and n is the preset number of nodes in each first hidden layer, where n is greater than or equal to x.

[0027] It is understandable that the more layers there are in the first hidden layer, the stronger the ability of the entire first preset neural network to analyze complex problems, and the number of elements allowed to be input in the first input layer, that is, the number of elements in the first vector, can also be appropriately increased. Based on this idea, this embodiment provides an improved method for determining the number of first hidden layers. The significance of the above formula is to control the total number of first hidden layers. When the number of input elements is small, the number of first hidden layers should be significantly increased to improve accuracy. When the number of input elements is large, the growth rate of the number of first hidden layers should be appropriately slowed down to prevent the amount of data from being too large. In addition, this embodiment further considers the number of nodes in each first hidden layer, limits the overall scale of the first preset neural network, and avoids the problems of overfitting and poor generalization ability caused by too many parameters in the first hidden layer.

[0028] Furthermore, in a preferred embodiment, in the above-mentioned primary control module 130, the predicted flushing volume of the water purifier can be obtained by any existing method, such as using a linear function to calculate based on the predicted total water consumption, or establishing a flushing analysis model based on the actual structure of the water purifier, and then obtaining it based on the predicted total water consumption.

[0029] Furthermore, in a preferred embodiment, the second prediction module 140 specifically includes: A second vector establishment module is used to establish a second vector sequence based on the target user's water usage record and the IoT monitoring record, wherein the second vector sequence includes multiple second vectors, each second vector corresponds to a time period, and the multiple second vectors are arranged based on the corresponding time sequence. The second vectors are used to represent the water usage characteristics of the target user in the corresponding time period and the environmental characteristics monitored by the IoT in the corresponding time period, wherein the element values ​​in the second vector are used to represent the corresponding time period, the number of water usage times in the time period obtained based on the water usage record, and the IoT monitoring record in the time period; The second network calling module is used to input the second vector sequence into the second preset neural network to obtain the predicted water use frequency output by the second preset neural network.

[0030] Unlike the first prediction module, the second prediction module is used to predict water usage frequency, which is data with obvious time attributes. Therefore, in this embodiment, the same water usage records and IoT monitoring records are re-encoded into a second vector sequence containing time information, and then analyzed using a second preset neural network.

[0031] The second preset neural network can directly use a recurrent neural network to analyze the second vector sequence. However, in reality, different users have different water usage goals, and therefore the correlation between water usage habits and IoT monitoring records is also different. If the same recurrent neural network is used for prediction, inaccurate predictions may occur.

[0032] The present invention takes this problem into consideration and therefore further provides a preferred solution: specifically, in a preferred embodiment, the second preset neural network includes a second recurrent neural network and a third recurrent neural network, wherein the second recurrent neural network is connected to the third recurrent neural network, the second recurrent neural network is used to output associated hyperparameters according to the second vector sequence, the associated hyperparameters are used to describe the degree of correlation between the target user's water use records and the Internet of Things monitoring records, and the third recurrent neural network is used to output the predicted water use frequency according to the second vector sequence and the associated hyperparameters.

[0033] The above design analyzes the second vector sequence through the second recurrent neural network and the third recurrent neural network respectively, analyzes the correlation degree through the second preset neural network to obtain the correlation hyperparameters, and then analyzes the correlation hyperparameters through the third preset neural network to obtain more accurate prediction results.

[0034] Specifically, in a preferred embodiment, the second recurrent neural network includes a second input layer, at least one second hidden layer and a second output layer connected in sequence, wherein the number of nodes in the second input layer is the same as the number of elements in the second vector, each node in the second input layer is used to input an element in the second vector, each node in the second input layer is connected to a node in an adjacent second hidden layer, and each node in the second hidden layer is also connected to other nodes in the same layer, the second hidden layer is used for cyclic prediction, the number of cycles of the second hidden layer is the same as the number of second vectors in the second vector sequence, the second output layer includes a node, the node of the second output layer is connected to all nodes in the connected second hidden layer, and the node of the second output layer is used to output associated hyperparameters in the last cycle.

[0035] The third recurrent neural network includes a third input layer, a third hidden layer and a third output layer connected in sequence, wherein the number of nodes in the third input layer is the same as the number of elements in the second vector, each node in the third input layer is used to input an element of the second vector, each node in the third input layer is connected to a node in the third hidden layer, and each node in the third hidden layer is also connected to a node in the second output layer. The third hidden layer is used for cyclic prediction, the number of cycles of the third hidden layer is the same as the number of second vectors in the second vector sequence, the third output layer includes one node, the node of the third output layer is connected to all nodes in the third hidden layer, and the node of the third output layer is used to output the predicted water use frequency in the last cycle.

[0036] In this embodiment, multiple second hidden layers can improve analysis accuracy. Assuming the second recurrent neural network is relatively accurate, the third recurrent neural network can be appropriately simplified. Specifically, when multiple second hidden layers are present, as long as the associated hyperparameters are accurate, only one third hidden layer is required in the third recurrent neural network. This ensures that the amount of data in the second preset neural network is not excessive while maintaining overall operational efficiency.

[0037] In addition, in this embodiment, the associated hyperparameters output by the second recurrent neural network are used as preset parameters for reference by the hidden layer of the third recurrent neural network during the first cycle, so that the third recurrent neural network takes into account the degree of correlation between the target user's water usage records and the IoT monitoring records when performing analysis, so as to output more accurate results.

[0038] Figure 2 The following is a simplified schematic diagram of the second preset neural network in this embodiment. The number of nodes in the first input layer, the number and number of nodes in the first hidden layer, the number of nodes in the second input layer, and the number of nodes in the second hidden layer are all simplified representations, and the lines representing the recurrent functions of the first and second hidden layers are omitted.

[0039] Furthermore, in a preferred embodiment, the above-mentioned secondary control module 150 is mainly used to adjust the flushing interval and flushing duration of the water purifier based on the predicted flushing volume of the water purifier, so that the water purifier can meet the specific usage habits of the target user. Specifically, when the predicted total amount of water consumption is the same, when the target user's water use frequency is high, it indicates that the user has the habit of using water in small amounts multiple times. At this time, the flushing operation strategy of the water purifier should also be changed accordingly, that is, flushing is performed multiple times in a short time to avoid the user waiting for a long time when using water. Conversely, when the predicted total amount of water consumption is the same, if the user's water use frequency is low, it indicates that the user has the habit of using water in small times and large amounts. At this time, the water purifier can be adjusted to flush less frequently and for longer periods of time, so as to ensure the cleanliness of the water purifier when the flushing function is turned on less frequently.

[0040] The present invention provides a water purifier control system based on the Internet of Things. The control system first obtains the water consumption records and Internet of Things monitoring records of a target user through a data acquisition module, then predicts the total water consumption of the target user within a target time period based on the water consumption records and Internet of Things monitoring records of the target user through a first prediction module, then obtains the predicted water purifier flushing volume of the target user within the target time period based on the predicted total water consumption through a primary control module, then predicts the predicted water consumption frequency of the target user within the target time period based on the water consumption records and Internet of Things monitoring records of the target user through a second prediction module, and finally adjusts the flushing interval and flushing duration of the water purifier based on the predicted water consumption frequency and the predicted water purifier flushing volume according to the predicted water consumption frequency, wherein the flushing interval is inversely proportional to the water consumption frequency, and the flushing duration is inversely proportional to the water consumption frequency. Compared with the prior art, the present invention combines the Internet of Things technology, on the one hand, uses the water consumption records and Internet of Things monitoring records to predict the total water consumption, thereby realizing the adjustment of the flushing volume, and on the other hand, uses the water consumption records and Internet of Things monitoring records to predict the water consumption frequency again, so that when the flushing volume is determined, the flushing interval and flushing duration can be more finely adjusted, thereby realizing the flexible change of the control strategy of the water purifier according to the user's work and rest habits, so that it can adapt to the user's usage habits and improve the user experience.

[0041] The above is a detailed description of an embodiment of the present invention. However, the content described is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A water purifier control system based on the Internet of Things, characterized in that: include: Data acquisition module, used to obtain water consumption records and IoT monitoring records of target users; The first prediction module is used to predict the total water consumption of the target user within the target time period based on the water consumption record of the target user and the IoT monitoring record; The primary control module is used to obtain the predicted flushing volume of the water purifier by the target user within the target time period based on the predicted total water consumption; The second prediction module is used to predict the target user's water usage frequency within the target time period based on the target user's water usage record and IoT monitoring record; The secondary control module is used to adjust the flushing interval and flushing time of the water purifier according to the predicted water usage frequency and the predicted flushing volume of the water purifier, wherein the flushing interval is inversely proportional to the water usage frequency, and the flushing time is inversely proportional to the water usage frequency.

2. The water purifier control system based on the Internet of Things according to claim 1 is characterized in that: The first prediction module includes: A first vector establishment module is configured to establish a first vector based on the target user's water consumption records and IoT monitoring records, wherein the first vector is used to represent the target user's water consumption characteristics and the environmental characteristics monitored by the IoT, wherein the element values ​​in the first vector are used to represent the historical total water consumption, the historical average water consumption change rate, and the various IoT monitoring records obtained based on the water consumption records; The first network calling module is used to input the first vector into the first preset neural network to obtain the predicted total water consumption output by the first preset neural network.

3. The water purifier control system based on the Internet of Things according to claim 2, characterized in that: The first preset neural network includes a first input layer, at least one first hidden layer and a first output layer connected in sequence, wherein the number of nodes in the first input layer is equal to the number of elements in the first vector, each node in the first input layer is used to input an element of the first vector, the number of first hidden layers is proportional to the number of elements in the first vector, and the number of nodes in the first output layer is one, which is used to output the predicted total water consumption.

4. The water purifier control system based on the Internet of Things according to claim 3 is characterized in that: The number of the first hidden layer is determined by the following formula: ； Wherein, N is the number of first hidden layers, [] is a rounding symbol, e is a natural constant, x is the number of elements in the first vector, a is a preset scaling factor, and n is the preset number of nodes in each first hidden layer, where n is greater than or equal to x.

5. The water purifier control system based on the Internet of Things according to claim 1, characterized in that: The second prediction module includes: A second vector establishment module is used to establish a second vector sequence based on the target user's water usage record and the IoT monitoring record, wherein the second vector sequence includes multiple second vectors, each second vector corresponds to a time period, and the multiple second vectors are arranged based on the corresponding time sequence. The second vectors are used to represent the water usage characteristics of the target user in the corresponding time period and the environmental characteristics monitored by the IoT in the corresponding time period, wherein the element values ​​in the second vector are used to represent the corresponding time period, the number of water usage times in the time period obtained based on the water usage record, and the IoT monitoring record in the time period; The second network calling module is used to input the second vector sequence into the second preset neural network to obtain the predicted water use frequency output by the second preset neural network.

6. The water purifier control system based on the Internet of Things according to claim 5, characterized in that: The second preset neural network includes a second recurrent neural network and a third recurrent neural network, wherein the second recurrent neural network is connected to the third recurrent neural network, the second recurrent neural network is used to output associated hyperparameters according to the second vector sequence, the associated hyperparameters are used to describe the degree of association between the target user's water usage records and the Internet of Things monitoring records, and the third recurrent neural network is used to predict the water usage frequency according to the second vector sequence and the associated hyperparameter output.

7. The water purifier control system based on the Internet of Things according to claim 6, characterized in that: The second recurrent neural network includes a second input layer, at least one second hidden layer and a second output layer connected in sequence, wherein the number of nodes in the second input layer is the same as the number of elements in the second vector, each node in the second input layer is used to input an element in the second vector, each node in the second input layer is connected to the node in the adjacent second hidden layer, and each node in the second hidden layer is also connected to other nodes in the same layer, the second hidden layer is used for cyclic prediction, the number of cycles of the second hidden layer is the same as the number of second vectors in the second vector sequence, the second output layer includes one node, the node of the second output layer is connected to all nodes in the connected second hidden layer, and the node of the second output layer is used to output associated hyperparameters in the last cycle.

8. The water purifier control system based on the Internet of Things according to claim 7, characterized in that: The third recurrent neural network includes a third input layer, a third hidden layer and a third output layer connected in sequence, wherein the number of nodes in the third input layer is the same as the number of elements in the second vector, each node in the third input layer is used to input an element of the second vector, each node in the third input layer is connected to a node in the third hidden layer, and each node in the third hidden layer is also connected to a node in the second output layer. The third hidden layer is used for cyclic prediction, the number of cycles of the third hidden layer is the same as the number of second vectors in the second vector sequence, the third output layer includes one node, the node of the third output layer is connected to all nodes in the third hidden layer, and the node of the third output layer is used to output the predicted water use frequency in the last cycle.

Images