Method and system for measuring sewage index

Abstract

The application provides a sewage index determination method, a sewage index determination system and a storage medium. The sewage index determination method comprises the following steps: obtaining historical data of sewage treatment and a query sample to be predicted; determining similar training samples corresponding to the query sample and associated input variables according to the historical data and the query sample; and performing slow feature regression analysis on the query sample according to the similar training samples and the associated input variables to determine the value of the sewage index of the query sample.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sewage treatment, and in particular to a sewage index determination method, a sewage index determination system, and a corresponding computer readable storage medium. BACKGROUND

[0002] China's per capita water resources is only one fourth of the global average, and it is a relatively poor country in terms of per capita water resources. However, with the continuous expansion of the city scale in China, a large number of population is constantly pouring in, and the use of water resources is gradually increasing, and the discharge of urban sewage is also increasing. In order to meet the requirements of sustainable development, improve the utilization rate of water resources, sewage treatment is becoming more and more important.

[0003] For the sewage treatment process, the accurate prediction of the key indicators of sewage treatment can effectively help the factory to judge and check whether the sewage treatment after treatment meets the treatment standard and other key problems, reduce the sewage treatment cost of the factory, and improve the sewage treatment efficiency. However, due to problems such as sensor failure, the historical data often has missing values, which greatly hinders the prediction of the key indicators of sewage. Soft measurement is an effective technology to solve the prediction of the key indicators of sewage. In the field of sewage treatment, due to the nonlinearity and dynamics of the sewage treatment process, common soft measurement methods such as support vector regression (SVR) are difficult to effectively solve the data missing problem, and the model prediction effect is poor, which is difficult to meet the actual demand.

[0004] In order to overcome the above-mentioned defects existing in the prior art, the present application provides a sewage index determination method, which can accurately predict the sewage index in the case of data missing, so as to determine the accurate value of the sewage index. SUMMARY

[0005] The following presents a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements of all aspects nor delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that is presented later.

[0006] In order to overcome the above-mentioned defects existing in the prior art, the present application provides a sewage index determination method, which can accurately predict the sewage index in the case of data missing, so as to determine the accurate value of the sewage index.

[0007] Specifically, the method for determining the pollutants according to the first aspect of the present application comprises the following steps: obtaining historical data of sewage treatment and a query sample to be predicted. According to the historical data and the query sample, similar training samples corresponding to the query sample and associated input variables are determined. According to the similar training samples and the associated input variables, a slow feature regression analysis is performed on the query sample to determine the value of the sewage index of the query sample.

[0008] Further, in some embodiments of the present application, the step of determining, according to the historical data and the query sample, similar training samples corresponding to the query sample and associated input variables comprises: checking the integrity of the query sample, and filling in the corresponding missing values if the query sample has data missing. According to the historical data, historical data with high similarity to the query sample are determined as the similar training samples. According to the historical data, candidate input variables with high correlation to the sewage index are determined as the associated input variables.

[0009] Further, in some embodiments of the present application, the step of checking the integrity of the query sample, and filling in the corresponding missing values if the query sample has data missing comprises: performing data preprocessing on the historical data. According to the preprocessed historical data, a latent posterior distribution of the historical data is determined. According to the historical data and the latent posterior distribution of the historical data, the data missing of the query sample is filled in.

[0010] Further, in some embodiments of the present application, the step of performing data preprocessing on the historical data comprises: performing mean-variance normalization on the historical data. According to the historical data after mean-variance normalization, the dimension of the latent variable in the historical data is determined.

[0011] Further, in some embodiments of the present application, the step of determining, according to the historical data, the historical data with high similarity to the query sample as the similar training samples comprises: determining a first SKL divergence of the historical data. According to the first SKL divergence, the difference degree of the latent variables of the historical data and the query sample is determined to select the similar training samples.

[0012] Further, in some embodiments of the present application, the step of determining, according to the first SKL divergence, the difference degree of the latent variables of the historical data and the query sample to select the similar training samples comprises: selecting a corresponding number N1 of historical data with smaller first SKL divergence from the historical data as the similar training samples according to a preset number N1.

[0013] Further, in some embodiments of the present application, the step of determining the candidate input variables with higher correlation with the sewage index as the associated input variables according to the historical data comprises: approximating the distribution of each of the candidate input variables and the predicted value by using a Gaussian kernel density estimation method. Calculating the probability distribution function of each of the candidate input variables. According to the distribution difference between the candidate input variables and the predicted value, determining the second SKL divergence of the candidate input variables. According to the second SKL divergence, determining the associated input variables from the candidate input variables.

[0014] Further, in some embodiments of the present application, the step of determining the candidate input variables with higher correlation with the sewage index as the associated input variables according to the second SKL divergence comprises: selecting a corresponding number N2 of candidate input variables with smaller second SKL divergence from the candidate input variables as the associated input variables according to a preset number N2.

[0015] Further, in some embodiments of the present application, the step of performing slow feature regression analysis on the query sample according to the similar training sample and the associated input variable to determine the value of the sewage index of the query sample comprises: determining a sample weight matrix according to the query sample and the similar training sample. According to the sample weight matrix, determining a weighted average matrix of the training sample input and output. According to the associated input variable, the similar training sample and the weighted average matrix, constructing and training a first slow feature regression analysis model. According to the first slow feature regression analysis model and the query sample, predicting the sewage index.

[0016] Further, in some embodiments of the present application, the step of predicting the sewage index according to the first slow feature regression analysis model and the query sample comprises: discarding a second slow feature regression analysis model that has been constructed or exists before the first slow feature regression analysis model, and inputting the query sample into the first slow feature regression analysis model to determine the value of the sewage index.

[0017] In addition, the above sewage index determination system according to the second aspect of the present application comprises a memory and a processor. The processor is connected to the memory and is configured to implement the above sewage index determination method according to the first aspect of the present application.

[0018] In addition, the above computer readable storage medium according to the third aspect of the present application has computer instructions stored thereon. When the computer instructions are executed by a processor, the above sewage index determination method according to the first aspect of the present application is implemented. BRIEF DESCRIPTION OF DRAWINGS

[0019] The above features and advantages of the present application will be better understood by reading the detailed description of the embodiments of the present application in conjunction with the drawings, in which:

[0020] Figure 1 A schematic diagram of a sewage index determination device according to some embodiments of the present application is shown;

[0021] Figure 2 A flow chart of a sewage index determination method according to some embodiments of the present application is shown;

[0022] Figure 3 A numerical diagram of predicted values and true values of a sewage index determination method according to some embodiments of the present application is shown;

[0023] Figure 4 A numerical diagram of predicted values and true values of a sewage index determination method according to some embodiments of the present application is shown;

[0024] Figure 5 A numerical diagram of predicted values and true values of a sewage index determination method according to some embodiments of the present application is shown. DETAILED DESCRIPTION

[0025] The specific embodiments of the present application will be described in the following detailed description, which should be considered in conjunction with the drawings. Although the description of the present application will be in conjunction with preferred embodiments, it is not meant to limit the present application to only these embodiments. On the contrary, the description of the present application is intended to cover all alternatives, modifications, and equivalent arrangements as can be included within the scope of the present application as defined by the appended claims. In order to provide an overall understanding of the application, numerous specific details are set forth in the following description. The present application can be practiced without some or all of these specific details. In other instances, well known structures and functions have not been described in detail in order to avoid obscuring the present application.

[0026] In the description of the present application, it is necessary to point out that, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through intermediate medium, can be internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0027] In addition, "upper", "lower", "left", "right", "top", "bottom", "horizontal", "vertical" used in the following description and shown in the drawings should be understood as the orientation shown in the drawing in question and related drawings. Such relative terms are used for convenience of explanation only and do not imply that the device described must be manufactured or operated in a particular orientation and should not be construed as limiting the present application.

[0028] It is to be understood that, although the terms "first", "second", "third", etc. can be used herein to describe various components, regions, layers and / or sections, these components, regions, layers and / or sections should not be limited by these terms and these terms are only used to distinguish different components, regions, layers and / or sections. Therefore, the first components, regions, layers and / or sections discussed below can be referred to as the second components, regions, layers and / or sections without departing from some embodiments of the present application.

[0029] As described above, for the sewage treatment process, accurate prediction of the key indicators of sewage treatment can effectively help the plant to effectively determine whether the sewage treatment after inspection meets the treatment standard and other key issues, reduce the cost of sewage treatment of the plant, and improve the efficiency of sewage treatment. However, due to sensor failure and other problems, missing data often exists in historical data, which causes great difficulty in predicting key indicators of sewage treatment. Soft measurement is an effective technology to solve the prediction of key indicators of sewage.

[0030] In order to overcome the above-mentioned defects existing in the prior art, the present application provides a sewage indicator measuring method, a sewage indicator measuring device and a corresponding computer readable storage medium, which are used to accurately predict the sewage indicator in the case of data missing to determine the accurate value of the sewage indicator.

[0031] In some non-limiting embodiments, the above-mentioned sewage indicator measuring method provided by the first aspect of the present application can be implemented by the above-mentioned sewage indicator measuring device provided by the second aspect of the present application. Specifically, the sewage indicator measuring device is configured with a memory and a processor. The memory includes but is not limited to the above-mentioned computer readable storage medium provided by the third aspect of the present application, and the computer instructions are stored on the computer readable storage medium. The processor is connected to the memory and is configured to execute the computer instructions stored on the memory to implement the sewage indicator measuring method provided by the first aspect of the present application.

[0032] First, please refer to Figure 1 , Figure 1 The architecture diagram of the pollutant measuring device according to some embodiments of the present application is shown.

[0033] Figure 1 An architecture diagram of the pollution determination device according to some embodiments of the present application is shown. The wastewater index determination device includes an internal communication bus 301, a processor 302, a read-only memory (ROM) 303, a random access memory (RAM) 304, a communication port 305, and a hard disk 307. The internal communication bus 301 can enable data communication between the components of the wastewater index determination device. The processor 302 can make judgments and issue prompts. In some embodiments, the processor 302 can be composed of one or more processors. The communication port 305 can enable data transmission and communication between the wastewater index determination device and external input / output devices. In some embodiments, the wastewater index determination device can send and receive information and data from a network through the communication port 305. In some embodiments, the wastewater index determination device can perform data transmission and communication with external input / output devices in a wired manner through the input / output port 306. The wastewater index determination device also includes different forms of program storage units and data storage units, such as the hard disk 307, the read-only memory (ROM) 303, and the random access memory (RAM) 304, which can store various data files used by the computer processing and / or communication, and possible program instructions executed by the processor 302. The processor 302 executes these instructions to implement the main part of the method. The results processed by the processor 302 are transmitted to external output devices through the communication port 305 and displayed on the user interface of the output devices.

[0034] The working principle of the wastewater index determination device described above will be described below in combination with some embodiments of the wastewater index determination method. Those skilled in the art can understand that the embodiments of the communication method are only some non-limiting embodiments provided by the present application, which are intended to clearly show the main idea of the present application and provide some specific schemes for facilitating the public to implement, rather than to limit the overall function or overall working mode of the wastewater index determination device. Similarly, the wastewater index determination device is also only a non-limiting embodiment provided by the present application, which does not limit the execution subject of each step in the wastewater index determination method.

[0035] Please refer to Figure 2 , Figure 2 A flowchart of the wastewater index determination method according to some embodiments of the present application is shown.

[0036] As shown in step S1 of Figure 2 In the process of determining the pollution, the wastewater index determination method can first obtain the historical data of wastewater treatment and the query sample to be predicted. Then, as shown in step S2 of Figure 2According to the historical data and the query sample, the present application can determine the similar training sample corresponding to the query sample and the associated input variable. Then, according to the similar training sample and the associated input variable, the present application can perform a slow feature regression analysis on the query sample to determine the value of the sewage index of the query sample. Figure 2 According to the historical data and the query sample, the present application can determine the similar training sample corresponding to the query sample and the associated input variable. Then, according to the similar training sample and the associated input variable, the present application can perform a slow feature regression analysis on the query sample to determine the value of the sewage index of the query sample.

[0037] Alternatively, in some embodiments of the present application, those skilled in the art can install sensors for measuring the above-mentioned sewage index to obtain the data of sewage treatment, and establish a sewage index database containing the historical data.

[0038] Those skilled in the art can understand that the above-mentioned sewage index sensor is only a non-limiting embodiment provided by the present application, which is intended to measure the historical index of the above-mentioned sewage treatment to obtain the historical data of the sewage index, and is not intended to limit the protection scope of the present application.

[0039] Alternatively, in some embodiments of the present application, the obtained historical data of sewage treatment can include but is not limited to one or more of biochemical oxygen demand (BOD5), dissolved oxygen content, chemical oxygen demand, total suspended solid concentration, water environmental quality index, dissolved readily biodegradable organic matter concentration, readily degradable substrate concentration, NO3 - N and NO2 - N concentration, NH4 - N and NH3 - N concentration, dissolved biodegradable organic nitrogen concentration, and readily degradable substrate concentration.

[0040] Further, the present application can determine the similar training sample corresponding to the query sample and the associated input variable according to the historical data and the query sample. Specifically, the present application can check the integrity of the query sample, and fill in the corresponding missing values if the query sample has data missing. According to the historical data, the historical data with high similarity to the query sample is determined as the similar training sample. According to the historical data, the candidate input variable with high correlation to the sewage index is determined as the associated input variable.

[0041] Further, the present application can check the integrity of the query sample. If the query sample has data missing, the corresponding missing values are filled in. Specifically, the historical data is preprocessed. According to the preprocessed historical data, the potential posterior distribution of the historical data is determined. According to the historical data and the potential posterior distribution of the historical data, the data missing of the query sample is filled in.

[0042] Preferably, in some embodiments of the present application, the step of data preprocessing the historical data can be mean-variance normalization of the historical data, and according to the mean-variance normalized historical data, the dimension of the latent variable in the historical data is determined.

[0043] Those skilled in the art can understand that the above-mentioned mean-variance normalization method is only a preferred solution provided by the present application, which is intended to preprocess the data to determine the dimension of the latent variable in the historical data, and is not used to limit the protection scope of the present application.

[0044] Specifically, the present application can first normalize the historical data by mean-variance, and the formula is as follows:

[0045]

[0046] Wherein z represents the standardized data, x represents the original data, μ represents the expectation of the original data, and σ represents the variance of the original data, the purpose is to normalize the current data by mean-variance, and reduce the influence of the latent variable with large variance. After that, principal component analysis [1] (hereinafter referred to as PCA) is used to determine the dimension of the latent variable to reduce the data processing time later.

[0047] After that, the present application can obtain the latent posterior distribution of the historical data through VBPCA (penalized probabilistic principal component analysis), and the main steps are as follows: VBPCA is based on probabilistic principal component analysis (hereinafter referred to as PPCA), and the parameter value of VBPCA is penalized. VBPCA assumes that the parameters of PPCA have a Gaussian prior distribution.

[0048] p(μ)=N(μ|0,ε μ I)

[0049]

[0050] Wherein W∈R D×K represents the loading matrix, W: represents the jth column of the loading matrix, represents the set of probabilistic model hyperparameters of VBPCA, θ={W, μ, e j represents the model parameters.

[0051] In the above steps, the analytical solution of the latent variable posterior distribution p9θ∣x j ,ε) cannot be obtained, in order to solve the problem, the variational Bayesian EM algorithm is introduced, the mean field approximation is used to simplify the posterior estimation process, and the coordinate ascent strategy is used to iteratively update the posterior distribution of unknown parameters, and the posterior distribution of the parameters is factorized as follows:

[0052]

[0053]

[0054] where x j ∈R D×1 represents the jth observation, z j ∈R K×1 represents the latent variable of the jth sample, W i : represents the ith column of the loading matrix, u i represents the mean of the ith sample, C(q(θ),ε) represents the cost function, and the probability density function q(θ) represents the best approximation form of the posterior distribution p(θ∣x j ,ε).

[0055] Thus, the present application can obtain the latent posterior distribution of the historical data by the above method.

[0056] Further, the present application can determine the difference degree of the latent variable of the query sample in the historical data according to the first SKL divergence of the historical data, so as to select the similar training sample.

[0057] Further, the present application can select the corresponding number N1 of historical data with smaller first SKL divergence from the historical data as the similar training sample according to the preset number N1.

[0058] Specifically, the present application assumes that the measurement noise is white Gaussian when selecting the sample and variable, and adopts the similarity measurement criterion of Gaussian distribution. The criterion uses the SKL divergence to calculate the difference degree of Gaussian distribution. The SKL divergence can be defined as follows:

[0059]

[0060] where x j: represents the jth historical sample, x query represents the query sample, trace represents the trace operator, z query represents the latent variable of the query sample, and are the mean and covariance value thereof, respectively.

[0061] The sample with smaller SKL divergence value is more similar to the query sample. The historical data will be arranged in ascending order according to the SKL divergence value, and the first N1 most similar samples will be selected as the above similar training sample according to experience for subsequent correlation analysis of the variable.

[0062] Further, the application can determine candidate input variables with high correlation with the sewage indicators as the associated input variables according to the historical data. Specifically, the application can approximate the distribution of each candidate input variable and the predicted value by using the Gaussian kernel density estimation method, calculate the probability distribution function of each candidate input variable, and determine the second SKL divergence of the candidate input variable according to the distribution difference between the candidate input variable and the predicted value. The associated input variable is determined from the candidate input variable according to the second SKL divergence.

[0063] Specifically, under the instant learning framework, after selecting the sample, appropriate variable selection based on the selected sample is also needed, and the main steps are as follows:

[0064] First, the distribution of each input variable and the predicted value is approximated by using the Gaussian kernel density estimation method. The formula is as follows:

[0065]

[0066]

[0067]

[0068] where M represents the length of the current database, X : represents the jth candidate input variable, X i,j represents the jth candidate input variable of the ith sample, Y represents the predicted vector, and h represents the bandwidth.

[0069] After calculating the probability distribution function of each candidate input variable, the distribution difference between the candidate input variable and the predicted value is analyzed by using the SKL divergence, and the formula is as follows:

[0070]

[0071] where P(X :j ) and P(Y) represent the probability distribution function of the jth candidate input variable and the predicted value.

[0072] Further, the application can select the corresponding number N2 of candidate input variables with smaller second SKL divergence from the candidate input variables as the associated input variables according to the preset number N2.

[0073] Specifically, in some embodiments of the application, the variable with smaller SKL divergence value can be regarded as the input variable related to the predicted value. All variables will be reordered in ascending order according to the SKL divergence value, and the first N2 most relevant variables are selected as the associated input variables according to experience.

[0074] Further, after determining the associated input variables and the similar training samples, the present application can perform a slow feature regression analysis on the query sample according to the associated input variables and the training samples to determine the value of the sewage index of the query sample. The present application can first determine a sample weight matrix according to the query sample and the similar training samples. Then, the present application can determine a weighted average matrix of the training sample input and output according to the sample weight matrix. Then, the present application can construct and train a first slow feature regression analysis model according to the associated input variables, the similar training samples and the weighted average matrix. Then, the present application can predict the sewage index according to the first slow feature regression analysis model and the query sample.

[0075] Further, in some embodiments of the present application, the present application can first discard a second slow feature regression analysis model that has been constructed or exists before the first slow feature regression analysis model. Then, the present application can input the query sample into the first slow feature regression analysis model that has been constructed to determine the value of the sewage index.

[0076] Specifically, the query sample X query is input into the first slow feature regression analysis model, and the similarity matrix Ψ between the query sample X selected and the training samples X

[0077]

[0078] where Ψ i (i = 1, 2, …, N1) is the similarity between the query sample X query and the ith training sample X , i.e., the sample weight. The ith diagonal element is defined according to the corresponding latent distribution dissimilarity, which can be calculated as follows:

[0079]

[0080] where std(·) calculates the standard deviation, and η represents an adjustable parameter. It is worth noting that when η→+∞, the weight of the selected sample approaches 1,

[0081] The weighted average matrix of the training sample input and output (X w , Y w ) is calculated, and the expression is as follows:

[0082]

[0083]

[0084] where, represents the ith training sample, BOD5 value of the ith training sample. The weighted mean of all variables of the training sample and the query sample need to be removed before constructing the locally weighted slow feature analysis model.

[0085]

[0086]

[0087]

[0088] where is an N1x1 matrix with all elements equal to 1, and N1 represents the number of training samples.

[0089] After scaling the selected input data and the query sample, the training samples will be weighted. Then, the basic linear slow feature analysis algorithm will be implemented on the weighted samples:

[0090]

[0091] where two-step singular value decomposition (SVD) is used. First, the covariance matrix of the input data is calculated as follows

[0092]

[0093] By singular value decomposition, the correlation between input variables can be eliminated

[0094] S = UDU T

[0095] Then the whitening matrix Q is obtained

[0096] Q = D -1 / 2 U T

[0097] In addition, the whitening transformation is given

[0098]

[0099]

[0100] where Using a similar method, another covariance

[0101]

[0102] where is the first-order derivative of Z. Next, the orthogonal matrix P can be derived by the following SVD

[0103]

[0104] The weighting matrix W can be calculated as follows

[0105]

[0106] Finally, the slow feature can be written as follows

[0107]

[0108] The coefficient θ between the slow feature and the output value is calculated. The relationship between the latent variable and the output value can be expressed in the form of weighted linear regression, and the formula is as follows:

[0109]

[0110] After the query sample is predicted, the predicted output value can be calculated as the predicted value of the above sewage index.

[0111]

[0112]

[0113]

[0114] After the query sample is updated, that is, the next round of sewage sample index determination is performed, the above-mentioned local prediction model is discarded, and the above-mentioned method is repeated to predict the sewage index of the updated query sample. In this way, the model can accurately predict the sewage index in the case of data missing to determine the accurate value of the sewage index.

[0115] Please further refer to Figure 3 , Figure 4 and Figure 5 . Figure 3 , Figure 4 , Figure 5 The numerical value diagram of the predicted value and the true value of the sewage index determination method provided by some embodiments of the present application is shown.

[0116] As shown in the figure, the deviation between the true value and the predicted value predicted by the above-mentioned sewage index determination method is very small, thereby illustrating that the model can accurately predict the sewage index in the case of data missing to determine the accurate value of the sewage index.

[0117] Although the above-mentioned method is illustrated and described as a series of actions for the purpose of simplifying the explanation, it should be understood and appreciated that the method is not limited by the order of actions, because according to one or more embodiments, some actions can occur in different order and / or concurrently with other actions illustrated and described herein or not illustrated and described herein but can be understood by those skilled in the art.

[0118] Those of skill would further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans can implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

[0119] Although the determination system of the wastewater indicator described in the above embodiments can be implemented by a combination of software and hardware, it is understood that the determination system of the wastewater indicator can also be implemented in software or hardware. For hardware implementation, the determination system of the wastewater indicator can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described above and a selection of these components can be implemented together in an exclusive-OR (XOR) logic device. For software implementation, the determination system of the wastewater indicator can be implemented by a combination of procedures and functions in software modules running on a general purpose chip.

[0120] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In the alternative, the processor and the storage medium can reside as discrete components in a user terminal.

[0121] In one or more exemplary embodiments, the functions described can be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functions can be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media can be any available media that can be accessed by a computer. By way of example, and not limitation, such computer-readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. Also, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. Disk and disc, as used herein, includes compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and blu-ray disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0122] The previous description of the disclosure is provided to enable any persons skilled in the art to make or use the disclosure. Various modifications to the disclosure will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for measuring a sewage index, characterized by, The method comprises the following steps: obtaining historical data of sewage treatment and a query sample to be predicted; determining similar training samples and associated input variables corresponding to the query sample according to the historical data and the query sample; determining a sample weight matrix according to the query sample and the similar training samples; determining a weighted average matrix of the training sample input and output according to the sample weight matrix; constructing and training a first slow feature regression analysis model according to the associated input variables, the similar training samples and the weighted average matrix; and predicting the sewage index according to the first slow feature regression analysis model and the query sample.

2. The assay method according to claim 1, characterized by The step of determining similar training samples and associated input variables corresponding to the query sample according to the historical data and the query sample comprises: checking the integrity of the query sample, and filling in the corresponding missing values if the query sample has data missing; determining historical data with high similarity to the query sample as the similar training samples according to the historical data; and determining candidate input variables with high correlation to the sewage index as the associated input variables according to the historical data.

3. The assay method according to claim 2, characterized in that, The step of checking the integrity of the query sample, and filling in the corresponding missing values if the query sample has data missing comprises: performing data preprocessing on the historical data; determining the latent posterior distribution of the historical data according to the preprocessed historical data; and filling in the data missing of the query sample according to the historical data and the latent posterior distribution of the historical data.

4. The assay method according to claim 3, characterized in that The step of performing data preprocessing on the historical data comprises: performing mean-variance normalization on the historical data; and determining the dimension of the latent variables in the historical data according to the historical data after mean-variance normalization.

5. The assay method according to claim 4, characterized in that The step of determining the historical data with high similarity to the query sample as the similar training samples according to the historical data comprises: determining the latent variables of the historical data and the first SKL divergence of the historical data; and determining the difference degree of the latent variables of the historical data and the query sample according to the first SKL divergence, so as to select the similar training samples.

6. The assay method according to claim 5, characterized in that, The step of determining the difference degree of the latent variables of the historical data and the query sample according to the first SKL divergence, so as to select the similar training samples comprises: selecting a corresponding number N1 of historical data with smaller first SKL divergence from the historical data as the similar training samples according to a preset number N1.

7. The assay method according to claim 2, characterized by, The step of determining candidate input variables with high correlation to the sewage index as the associated input variables according to the historical data comprises: approximating the distribution of each candidate input variable and predicted value by a Gaussian kernel density estimation method; calculating the probability distribution function of each candidate input variable; determining the second SKL divergence of the candidate input variables according to the distribution difference between the candidate input variables and the predicted value; and determining the associated input variables from the candidate input variables according to the second SKL divergence.

8. The assay method according to claim 7, characterized by The step of determining the associated input variables from the candidate input variables according to the second SKL divergence comprises: selecting a corresponding number N2 of candidate input variables with smaller second SKL divergence from the candidate input variables according to a preset number N2 as the associated input variables.

9. The assay method of claim 1, wherein The step of predicting the sewage index according to the first slow feature regression analysis model and the query sample comprises: discarding a second slow feature regression analysis model that has been constructed or exists before the first slow feature regression analysis model, and inputting the query sample into the first slow feature regression analysis model to determine the value of the sewage index.

10. A sewage index determination system, comprising: a memory; and a processor connected to the memory and configured to implement the sewage index determination method according to any one of claims 1-9.

11. A computer readable storage medium having stored thereon computer instructions, wherein, The computer instructions are executed by the processor to implement the sewage index determination method according to any one of claims 1-9.

Images