CN111047474A - Indoor harmful substance volatilization time estimation method and device and storage medium - Google Patents
Indoor harmful substance volatilization time estimation method and device and storage medium Download PDFInfo
- Publication number
- CN111047474A CN111047474A CN201911001463.9A CN201911001463A CN111047474A CN 111047474 A CN111047474 A CN 111047474A CN 201911001463 A CN201911001463 A CN 201911001463A CN 111047474 A CN111047474 A CN 111047474A
- Authority
- CN
- China
- Prior art keywords
- volatilization
- data information
- neural network
- related data
- network model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 239000000126 substance Substances 0.000 title claims abstract description 63
- 238000003062 neural network model Methods 0.000 claims abstract description 146
- 239000002245 particle Substances 0.000 claims abstract description 98
- 238000012549 training Methods 0.000 claims description 63
- 239000013598 vector Substances 0.000 claims description 35
- 238000004364 calculation method Methods 0.000 claims description 34
- 238000012545 processing Methods 0.000 claims description 23
- 239000000463 material Substances 0.000 claims description 19
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 4
- 238000009423 ventilation Methods 0.000 description 10
- WSFSSNUMVMOOMR-UHFFFAOYSA-N Formaldehyde Chemical compound O=C WSFSSNUMVMOOMR-UHFFFAOYSA-N 0.000 description 9
- 238000005034 decoration Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 9
- 210000002569 neuron Anatomy 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- 239000007787 solid Substances 0.000 description 5
- 239000003973 paint Substances 0.000 description 4
- 239000002023 wood Substances 0.000 description 4
- UHOVQNZJYSORNB-UHFFFAOYSA-N Benzene Chemical compound C1=CC=CC=C1 UHOVQNZJYSORNB-UHFFFAOYSA-N 0.000 description 3
- 231100000614 poison Toxicity 0.000 description 3
- OIGNJSKKLXVSLS-VWUMJDOOSA-N prednisolone Chemical compound O=C1C=C[C@]2(C)[C@H]3[C@@H](O)C[C@](C)([C@@](CC4)(O)C(=O)CO)[C@@H]4[C@@H]3CCC2=C1 OIGNJSKKLXVSLS-VWUMJDOOSA-N 0.000 description 3
- 238000013139 quantization Methods 0.000 description 3
- 239000003440 toxic substance Substances 0.000 description 3
- 239000011248 coating agent Substances 0.000 description 2
- 238000000576 coating method Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 241000345998 Calamus manan Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- -1 artificial board Substances 0.000 description 1
- 235000013339 cereals Nutrition 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 230000008020 evaporation Effects 0.000 description 1
- 238000001704 evaporation Methods 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 210000001145 finger joint Anatomy 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- WSFSSNUMVMOOMR-NJFSPNSNSA-N methanone Chemical compound O=[14CH2] WSFSSNUMVMOOMR-NJFSPNSNSA-N 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000011120 plywood Substances 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 229910052704 radon Inorganic materials 0.000 description 1
- SYUHGPGVQRZVTB-UHFFFAOYSA-N radon atom Chemical compound [Rn] SYUHGPGVQRZVTB-UHFFFAOYSA-N 0.000 description 1
- 235000012950 rattan cane Nutrition 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Air Conditioning Control Device (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses an indoor harmful substance volatilization time estimation method, device and storage medium, specifically comprising: acquiring volatilization related data information provided by a client, wherein the volatilization related data information is data information representing volatilization time of harmful substances influencing the indoor environment; inputting the acquired volatilization related data information to a reverse relay Broadcasting (BP) neural network model trained in advance, wherein the BP neural network model adopts a variable weight particle swarm method for convergence; and the BP neural network model calculates volatilization time according to the input volatilization related data information, and returns the volatilization time to the client as an estimation result of the volatilization time of the indoor harmful substances. By applying the scheme of the embodiment of the application, the volatilization time of indoor harmful substances can be accurately and conveniently obtained without manual work or instrument detection of a user, and the personal safety of the user is more effectively ensured.
Description
Technical Field
The application relates to the technical field of computers, in particular to a method and a device for estimating volatilization time of indoor harmful substances and a storage medium.
Background
Interior decoration is more and more common in people's life, and more or less harmful substance such as formaldehyde is usually contained in the decoration material. Due to inhalation of human bodyThe amount of harmful substances can easily cause body diseases, so that the industry has specified a unified standard for the content of the harmful substances, such as a formaldehyde concentration standard of 0.1mg/m3. It is safe to live in when the concentration of the indoor harmful substances is lower than the unified standard. However, in actual life, not only are the types of interior materials varied, but also the volatilization of harmful substances is related to various other factors, such as the indoor area, the ventilation condition and the like, and it is difficult to clearly determine the volatilization time of the harmful substances until the harmful substances are volatilized below the safe concentration.
Disclosure of Invention
Aiming at the prior art, the embodiment of the invention discloses a method for estimating volatilization time of indoor harmful substances, which can overcome the defects that the volatilization time of indoor toxic substances is inconvenient and inaccurate for a user to estimate and achieve the aim of conveniently and accurately estimating the volatilization time of the indoor toxic substances.
The method for estimating volatilization time of indoor harmful substances provided by the embodiment of the application specifically comprises the following steps:
acquiring volatilization related data information provided by a client, wherein the volatilization related data information is data information representing volatilization time of harmful substances influencing the indoor environment;
inputting the acquired volatilization related data information to a reverse relay Broadcasting (BP) neural network model trained in advance, wherein the BP neural network model adopts a variable weight particle swarm method for convergence;
and the BP neural network model calculates volatilization time according to the input volatilization related data information, and returns the volatilization time to the client as an estimation result of the volatilization time of the indoor harmful substances.
Further, the air conditioner is provided with a fan,
the method for calculating the volatilization time by the BP neural network model according to the input volatilization related data information comprises the following steps:
the input layer of the BP neural network model receives the volatilization related data information and directly transmits the received volatilization related data information to the hidden layer;
the hidden layer calculates the volatilization time according to the volatilization related data information and the set model parameters, and transmits the calculated volatilization time to the output layer;
and the output layer outputs the calculated volatilization time.
Further, the air conditioner is provided with a fan,
before the step of acquiring volatilization related data information, the method further comprises a step of training a BP neural network model, wherein the step of training the BP neural network model comprises the following steps:
acquiring a training set containing multiple groups of volatilization related data information;
initializing the parameters of the BP neural network model;
taking a group of volatilization related data information in the training set as current volatilization related data information, inputting the current volatilization related data information to an input layer of the BP neural network model, and directly transmitting the current volatilization related data information to a hidden layer of the BP neural network model by the input layer;
the hidden layer calculates to obtain a training value according to the current volatilization related data information and the parameters of the BP neural network model, and outputs the calculated training value through an output layer of the BP neural network model;
according to the output training value, correcting the parameters of the BP neural network model by using a variable weight particle swarm method;
and when the BP neural network model is not converged, taking the next group of volatilization related data information in the training set as the current volatilization related data information, and returning the step of inputting the current volatilization related data information to the input layer of the BP neural network model to be repeatedly executed until the BP neural network model is converged.
Further, the air conditioner is provided with a fan,
the method for correcting the parameters of the BP neural network model by using the variable weight particle swarm method comprises the following steps:
taking the parameters of the BP neural network model as the current position vectors of the particles in the variable weight particle swarm method;
calculating the update speed of the particles by using the current position vector of the particles, the set particle motion speed and the variable motion inertia weight;
calculating an updated position vector of a particle using the updated velocity of the particle and the current position vector of the particle;
and converting the updated position vector to be used as a parameter of the modified BP neural network model.
The embodiment of the application also provides an indoor harmful substance volatilization time estimation device, which can overcome the defects that the estimation of the volatilization time of indoor toxic substances by a user is inconvenient and inaccurate, and achieve the purpose of conveniently and accurately estimating the volatilization time.
The utility model provides an indoor harmful substance evaporation time estimation device, the device includes:
the system comprises a receiving unit, a control unit and a processing unit, wherein the receiving unit is used for acquiring volatilization related data information provided by a client, and the volatilization related data information is data information representing volatilization time of harmful substances influencing indoor;
the estimation unit is used for inputting the acquired volatilization related data information to a reverse relay propagation (BP) neural network model trained in advance, and the BP neural network model adopts a variable weight particle swarm method for convergence; calculating volatilization time according to the input volatilization related data information;
and the sending unit is used for taking the volatilization time as an estimation result of the volatilization time of the indoor harmful substances and returning the estimation result to the client.
Further, the air conditioner is provided with a fan,
the estimation unit includes:
the input layer computing unit is used for receiving the volatilization related data information and directly transmitting the received volatilization related data information to the hidden layer computing unit;
the hidden layer calculation unit is used for calculating the volatilization time according to the volatilization related data information and the set model parameters and transmitting the calculated volatilization time to the output layer calculation unit;
and the output layer calculating unit is used for outputting the calculated volatilization time.
Further, the air conditioner is provided with a fan,
the apparatus further comprises:
the model training processing unit is used for acquiring a training set containing multiple groups of volatilization related data information; taking a group of volatilization related data information in the training set as current volatilization related data information, and inputting the current volatilization related data information to an input layer of the BP neural network model; when the BP neural network model is not converged, taking the next group of volatilization related data information in the training set as the current volatilization related data information, and returning the step of inputting the current volatilization related data information to the input layer of the BP neural network model to be repeatedly executed until the BP neural network model is converged;
and the parameter processing unit is used for initializing the parameters of the BP neural network model and correcting the parameters of the BP neural network model by using a variable weight particle swarm method according to the output training value.
Further, the air conditioner is provided with a fan,
the parameter processing unit includes:
the initialization unit is used for initializing the parameters of the BP neural network model;
and the correcting unit corrects the parameters of the BP neural network model by using a variable weight particle swarm method according to the output training value.
Further, the air conditioner is provided with a fan,
the correction unit includes:
the particle position setting unit is used for taking the parameters of the BP neural network model as the current position vectors of the particles in the variable weight particle swarm method;
the particle motion calculation unit is used for calculating the update speed of the particles by using the current position vector of the particles, the set particle motion speed and the variable motion inertia weight; calculating an updated position vector of a particle using the updated velocity of the particle and the current position vector of the particle;
and the conversion unit is used for converting the updated position vector as a parameter of the corrected BP neural network model.
The embodiment of the present application also provides a computer readable storage medium, on which computer instructions are stored, wherein the instructions, when executed by a processor, can implement the steps in the above method for estimating volatilization time of indoor harmful substances.
In summary, the embodiments of the present application provide a method and an apparatus for estimating volatilization time of an indoor harmful substance, and a storage medium. In the scheme of the embodiment of the application, a BP neural network model is established through a large amount of volatilization related data information, and the BP neural network model is converged by adopting a variable weight particle swarm method, so that when a user provides actual volatilization related data information, the volatilization related data information can be transmitted to a server side from a client side, and the volatilization time can be directly calculated by the server side by utilizing the estimation of the established BP neural network model and is returned to the client side. Therefore, the user of the client can conveniently and accurately obtain the volatilization time of the indoor harmful substances without manual work or instrument detection.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive labor.
Fig. 1 is a schematic diagram of an application scenario 100 according to an embodiment of the present application.
Fig. 2 is a flowchart of a method 200 for estimating volatilization time of indoor harmful substances according to an embodiment of the present application.
Fig. 3 is a schematic diagram of a BP neural network model 300 that needs to be trained according to an embodiment of the present application.
Fig. 4 is a flowchart of a method 400 for training a BP neural network model according to an embodiment of the present application.
Fig. 5 is a flowchart of a method 500 for estimating volatilization time of indoor harmful substances according to an embodiment of the present application.
Fig. 6 is a block diagram of an apparatus 600 for estimating a volatilization time of a harmful substance in a room according to an embodiment of the present application.
Fig. 7 is a block diagram of an apparatus 700 for estimating a time of volatilization of a harmful substance in a room according to an embodiment of the present application.
Fig. 8 is a schematic diagram of an internal structure 800 of the parameter processing unit 605 according to the embodiment of the present application.
Fig. 9 is a schematic diagram of a terminal device 900 according to another embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims, as well as in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are, for example, capable of operation in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprising" and "having," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements explicitly listed, but may include other steps or elements not explicitly listed or inherent to such process, method, article, or apparatus.
The technical solution of the present invention will be described in detail with specific examples. Several of the following embodiments may be combined with each other and some details of the same or similar concepts or processes may not be repeated in some embodiments.
In order to conveniently and accurately estimate the volatilization time of the indoor harmful substances, the embodiment of the application takes the data information which influences the volatilization time of the indoor harmful substances as the input of the BP neural network model, and the volatilization time is automatically estimated by utilizing the calculation of the BP neural network model, so that the volatilization time of the indoor harmful substances can be conveniently and accurately provided for a user. In addition, the BP neural network model utilized in the embodiment of the application adopts a variable weight particle swarm method to converge, so that the convergence speed of the BP neural network model is higher.
Fig. 1 is a schematic diagram of an application scenario 100 according to an embodiment of the present application. As shown in fig. 1, the scenario includes a server 101 and several clients 102. The client 102 may provide the volatilization-related data information to the server 101, and the server 101 estimates the volatilization time by using the BP neural network model and returns the calculated volatilization time to the client 102. Thus, the user of the client 102 can confirm the check-in time according to the obtained volatilization time, so that the harm of indoor harmful substances to the body is avoided, and the personal safety is ensured.
Fig. 2 is a flowchart of a method 200 for estimating volatilization time of indoor harmful substances according to an embodiment of the present application. As shown in fig. 2, the method includes:
step S1: and acquiring volatilization related data information provided by a client, wherein the volatilization related data information is data information representing volatilization time of harmful substances influencing the indoor environment.
The harmful substances in the indoor decoration can be any pollutant in indoor decoration, such as radon, formaldehyde, benzene and the like. The volatilization related data information is data information which influences volatilization time of indoor harmful substances, such as total amount of paint used, floor material, furniture material, whether ventilation is needed, finishing finish time and the like.
The data is information related to the volatilization time of the indoor harmful substances, the prior art usually adopts manual estimation, the estimation result is often unreliable, or the field detection is realized by using an instrument, but the labor and the financial resources are wasted. Aiming at the problem of nonlinearity, the expression is difficult to be carried out by using a simple function model, and the embodiment of the application adopts a BP neural network model to simulate the real situation, so that the volatilization time can be conveniently and accurately estimated.
Step S2: and inputting the acquired volatilization related data information to a reverse relay (BP) neural network model trained in advance, wherein the BP neural network model adopts a variable weight particle swarm method for convergence.
The BP neural network model in the step is a multi-layer feedforward neural network trained according to an error back propagation algorithm, has arbitrary complex mode classification capability and good multi-bit function mapping capability, and comprises an input layer, a hidden layer and an output layer. It should be noted that, the present application does not adopt a gradient descent method to converge the BP neural network model, but proposes a variable weight particle swarm method to converge.
Step S3: and the BP neural network model calculates volatilization time according to the input volatilization related data information, and returns the volatilization time to the client as an estimation result of the volatilization time of the indoor harmful substances.
That is to say, in the embodiment of the application, volatilization-related data information is used as input of the BP neural network model, volatilization time is calculated by using the BP neural network model, and the calculated volatilization time is returned to the client as an estimation result of volatilization time of indoor harmful substances. The user of the client can accurately and conveniently obtain the volatilization time of the indoor harmful substances without using an instrument for detection, and the personal safety of the user is ensured to the greatest extent.
To better illustrate the solution of the present application, another example is described in detail below. The embodiment of the application needs to establish and train a BP neural network model in advance. Those skilled in the art will appreciate that training a neural network model requires a large number of input values in order to allow the model parameters to converge stably so that the output result values match the actual situation, at which point the training can be terminated.
It is assumed that "volatilization-related data information" in the embodiment of the present application is an input value of the BP neural network model, and "volatilization time" is an output value of the BP neural network model. The 'volatilization related data information' represents data information influencing volatilization time of indoor harmful substances, and comprises conditions such as house area, house direction, total coating usage amount, ventilation, floor material, furniture material, finishing time and the like. The "volatilization time" indicates a time required for the indoor harmful substance represented by the "volatilization-related data information" to volatilize to a safe concentration or less. In practical application, the BP neural network model between an input value and an output value can be used for acquiring a large amount of volatilization related data information and volatilization time, so that the volatilization time which is in line with practical conditions can be output after the practical volatilization related data information is input through calculation of the BP neural network model.
Fig. 3 is a schematic diagram of a BP neural network model requiring training according to an embodiment of the present application. As shown in fig. 3, the BP neural network model includes an input layer 301, a hidden layer 302, and an output layer 303. Fig. 4 is a flowchart of a method 400 for training a BP neural network model according to an embodiment of the present application. As shown in fig. 4, the method includes:
step T1: and acquiring a training set containing multiple groups of volatilization related data information.
In the embodiment of the application, a large number of input values are needed when the BP neural network is trained, and each group of input values is a group of volatilization related data information and is stored in the training set. Factors affecting the volatilization time of indoor harmful substances are assumed to include the area of a house, the direction of the house, the total amount of paint used, whether ventilation is available, the material of a floor, the material of furniture, the finishing time and the like. For the convenience of subsequent calculation, it is necessary to digitize the information such as "house direction", "ventilation", "floor material", and "furniture material" that is difficult to be represented by numerical values as follows:
assuming that "house Direction" is "Direction", including true south, true north, true east, true west, south east, south west, north east, and north west, "south", "not", "east", "west", "sout", "southeast", "not", and "not", respectively, the description thereof can be expressed by the following equation 1:
(formula 1) { south, normal, east, west, southeast, southwest, norathast, norathest } formula 1
After the quantization, the following formula 2 can be used to represent:
assuming that "ventilation" includes "ventilation" and "non-ventilation", which are "Wind" and "isfind", respectively, the numerical expression can be expressed as the following formula 3:
assuming that the "floor material" is "sMaterial", including density board, core board, plywood, multi-layer finger board, veneer board, wood finger board without formaldehyde glue, and cereal board, "DensityM", "coreM", "plyM", "multiM", "veneerM", "solidM", "vsolidM", "hexM", the description of which can be expressed by the following equation 4:
sMaterial [ { densityM, coreM, plyM, mulM, veneerM, solidM, vsolidM, hexM } equation 4
After the quantization, the following formula 5 can be used to represent:
assuming that the furniture material is "fMaterial", including solid wood, artificial board, metal, rattan plaited, and soft, respectively, "solid", "artificial W", "metalW", "ratanW", and "rheidW", the description thereof can be expressed by the following equation 6:
fMaterial { solid, artificailW, metalW, ratanW, rheidW } equation 6
After the quantization, the following formula 7 can be used to represent:
in addition, the "floor area", "total amount of paint used", and "finishing time" may be directly expressed by numerical values, or may be converted into other numerical values more convenient to calculate, and will not be described herein again.
Assuming that the factors affecting the volatilization time of the harmful substances in the room include "Area", "Direction of room", "total amount of paint used", "ventilation", "floor material", "furniture material" and "finishing time", which are described above, respectively, as "Area", "Direction", "Weight", "isopind", "sMaterial", "fMaterial" and "endTime", the "volatilization-related data information" can be described by the following equation 8, and the numerical result thereof is used as a set of volatilization-related data information Hk。
HkArctic, Direction, Weight, isWind, sMaterial, fMaterial, endTime equation 8
In practical application, this step will obtain several groups HkThese H arekForm training set H ═ H1,H2,H3,……Hn}。
And step T2, initializing the parameters of the BP neural network model.
In practical applications, there are a large number of parameters, such as connection weights and thresholds, in the BP neural network model, which are collectively referred to as parameters in the embodiments of the present application, and initialization processing is required for the parameters.
Step T3: and taking a group of volatilization related data information in the training set as the current volatilization related data information.
In the embodiment of the application, the volatilization related data information in the training set is input to the input layer of the BP neural network one by one until the model converges.
Step T4: and inputting the current volatilization related data information to an input layer of the BP neural network model, and directly transmitting the current volatilization related data information to a hidden layer of the BP neural network model by the input layer.
Assume that an input training sample is Hu∈{H1,H2,H3,……HnAfter passing through the input layer, the output O of a certain neuron iiCan be expressed as:
namely: output of neuron i OiAnd input
The same is true.
Step T5: and the hidden layer calculates to obtain a training value according to the current volatilization related data information and the parameters of the BP neural network model, and outputs the calculated training value through an output layer of the BP neural network model.
For the hidden and output layers, the output of neuron j is OjCan be expressed as:
Oj＝f(∑wjiOi-θj) Equation 10
Wherein f denotes an excitation function, wjiRepresents the connection weight, θ, of neuron i to neuron jjIs the threshold of neuron j, OiIs both the output of the previous layer of neurons i and the input of the next layer of neurons j.
The above steps T4 and T5 are the processing procedure of the BP neural network from the input layer to the output layer. It can be seen that, in the processing process, the parameters of the BP neural network model participate in the calculation, and finally the output training value is obtained. Here, since the BP neural network model is still in the training process, the output value is referred to as a training value.
Step T6: and correcting the parameters of the BP neural network model by using a variable weight particle swarm method according to the output training value.
Those skilled in the art will appreciate that the neural network model can be used only by training, and the internal parameters of the neural network model converge to a stable state. The BP neural network model in the embodiment of the present application also needs to continuously modify parameters in the training process, so that the model converges. The existing BP neural network model usually utilizes a gradient descent method, and the embodiment of the application adopts a variable weight particle swarm method to correct the parameters of the BP neural network model, so that the convergence speed can be accelerated.
The embodiment of the application discloses a method for correcting parameters of a BP neural network model by using a variable weight particle swarm method. The method is a method for searching the optimal position by assuming that m particles form a community in a D-dimensional target search space, each particle can be represented as a D-dimensional vector, and then the particle motion mode is utilized. In the embodiment of the present application, a parameter to be corrected is taken as a particle, and step T6 specifically includes:
step T61: and taking the parameters of the BP neural network model as the current position vector of the particles in the variable weight particle swarm method.
Here, the connection weight w in the BP neural network modeljiAnd a threshold value thetajIs a parameter, which is referred to herein as the position vector of the particle.
Step T62: and calculating the update speed of the particles by using the current position vector of the particles, the set motion speed of the particles and the variable motion inertia weight.
Step T63: calculating an updated position vector of the particle using the updated velocity of the particle and the current position vector of the particle.
The above steps T62 and T63 are methods of particle motion, which can be described as:
wherein m represents the m-th particle, k represents the particle motion at time k,
inertial weight, V, representing particle motionmRepresenting the velocity of the particle movement, EmA current position vector representing the particle is shown,
representing individual extrema in the search, gbpgRepresenting a global extremum in the search, C1And C2Denotes a learning factor, R1And R2All indicate thatRandom numbers uniformly distributed in the interval (0, 1). In addition, in the embodiments of the present application
The inertia weight can be updated after each movement of the particles, so that the local search capability and the global search capability are balanced, the convergence speed can be increased, and the purpose of falling into local optimum is avoided.
Step T64: and taking the updated position vector as a parameter of the modified BP neural network model.
That is, since parameters such as the connection weight w in the BP neural network are setjiAnd a threshold value thetajAs the position vector of the particles in the variable weight particle swarm method, the position vector can be continuously changed through the motion of the particles, namely, the parameters are changed or corrected, so that the convergence speed of the BP neural network model can be accelerated.
In practical application, the BP neural network model can tend to be stable and convergent only through multiple iterations. It can be generally measured by a fitness function:
wherein, FkFitness function, initime, representing time kiIndicating the true time of volatilization, inTimei' represents a training value of the network output, N represents the total number of iterations,
representing the sum of squared errors between the true volatilization time and the training value.
If FkIf the calculated output value is small enough, the calculated output value of the BP neural network model is close to the true value.
In practical application, after the m-th particle passes the k-th iteration, the inertia weight of the m-th particle
The variation of (d) can be expressed as:
ΔF＝Fk-Fk-1equation 13
Wherein, the delta F represents the variation value of the adaptive value, when the delta F is larger,
becoming relatively larger, the global search capability may be increased; in contrast, when Δ F is small,
becoming relatively small, the local search capability may be increased. Accordingly, in the embodiments of the present application
The method can balance the local search capability and the global search capability, accelerate the convergence speed and avoid falling into the aim of local optimum.
Step T7: and taking the next group of volatilization related data information in the training set as the current volatilization related data information, and returning to the step T4 to continue to repeat the execution until the BP neural network model converges.
In practical applications, the BP neural network model may be converged after several iterations, and therefore, the steps T4 to T7 may be repeatedly performed. After the BP neural network model is established by the scheme of the embodiment of the application, the volatilization related data information provided for the client side by the model can be used for calculating the volatilization time subsequently, so that the service is conveniently and accurately provided for the user.
Fig. 5 is a flowchart of a method 500 for estimating volatilization time of indoor harmful substances according to another embodiment of the present application, and the BP neural network model established as described above is used. As shown in fig. 5, the method includes:
step M1: and acquiring volatilization related data information provided by a client, wherein the volatilization related data information is data information representing volatilization time of harmful substances influencing the indoor environment.
This step is the same as step S1 of the above-described embodiment, that is, the client 102 may provide the volatilization-related data information obtained by itself to the server 101. Suppose the specific indoor decoration situation of a certain user is: the house area is 100 square meters, the house direction is south, the total amount of the coating is 50 kilograms, the house is in a ventilation state after decoration, the floor is made of solid wood finger joint plates, the furniture is made of solid wood, and the finishing time of decoration is 2019-05-01. According to the above-described embodiments, the numeralization method, the description of which can be expressed as:
“Area”＝100；“Direction”＝0；“Weight”＝50；“isWind”＝1，“sMaterial”＝5，“fMaterial”＝0，“endTime”＝20190501
then, the volatilization-related data information obtained in this step is {100,0,50,1,5,0,20190501 }.
Step M2: and inputting the acquired volatilization related data information to an input layer of a pre-trained BP neural network model.
Still taking the above assumption as an example, this step should input the set of data {100,0,50,1,5,0,20190501} to the input layer of the BP neural network model.
Step M3: and the input layer of the BP neural network model receives the volatilization related data information and directly transmits the received volatilization related data information to the hidden layer of the BP neural network model.
Step M4: and the hidden layer of the BP neural network model calculates the volatilization time according to the volatilization related data information and the set model parameters, and transmits the calculated volatilization time to the output layer.
Here, steps M3 and M4 are the main steps of BP neural network computation, which parameter connects the connection weights w between neuronsjiAnd a threshold value thetajThe parameters obtained when the BP neural network model is converged can be directly trained by using the embodiment. The specific calculations of the input layer, the hidden layer, and the output layer may refer to the above equations 9 and 10.
Step M5: and the output layer of the BP neural network model outputs the calculated volatilization time.
Through calculation, it is assumed that it takes 3 months to calculate the volatilization time of the harmful substances in the room. At this time, the step can directly return the result that the volatilization time is 3 months to the client, or return 20190801 time point to the client according to the decoration time 20190501, which indicates that the indoor harmful substances can be volatilized below the safe concentration until 2019-08-01.
The embodiment of the present application further provides an indoor harmful substance volatilization time estimation device 600, and fig. 6 is a schematic view of an internal structure of the device. As shown in fig. 6, the apparatus includes: receiving section 601, estimating section 602, and transmitting section 603. Wherein:
the receiving unit 601 is configured to obtain volatilization-related data information provided by the client, where the volatilization-related data information is data information indicating volatilization time of a harmful substance affecting a room.
An estimating unit 602, configured to input the obtained volatilization related data information to a BP neural network model trained in advance, where the BP neural network model adopts a variable weight particle swarm method to converge; and calculating the volatilization time according to the input volatilization related data information.
A sending unit 603, configured to return the volatilization time to the client as an estimation result of volatilization time of the indoor harmful substance.
That is, the receiving unit 601 acquires volatilization-related data information provided by the client; the estimation unit 602 inputs the acquired volatilization related data information to a BP neural network model trained in advance, and the BP neural network model adopts a variable weight particle swarm method for convergence; calculating volatilization time according to the input volatilization related data information; the sending unit 603 returns the volatilization time to the client as the estimation result of the volatilization time of the indoor harmful substance.
To better describe the above aspects of embodiments of the present application, another embodiment of the apparatus is described in detail below. Fig. 7 is a schematic structural diagram of another embodiment 700 of an apparatus according to an embodiment of the present application, and as shown in fig. 7, the apparatus further includes a model training processing unit 604 and a parameter processing unit 605 in addition to a receiving unit 601, an estimating unit 602, and a transmitting unit 603. Wherein:
a model training processing unit 604, configured to obtain a training set including multiple sets of volatilization related data information; taking a group of volatilization related data information in the training set as current volatilization related data information, and inputting the current volatilization related data information to an input layer of the BP neural network model; and when the BP neural network model is not converged, taking the next group of volatilization related data information in the training set as the current volatilization related data information, and returning the step of inputting the current volatilization related data information to the input layer of the BP neural network model to be repeatedly executed until the BP neural network model is converged.
A parameter processing unit 605, configured to perform initialization processing on the parameters of the BP neural network model, and modify the parameters of the BP neural network model according to the output training value and by using a variable weight particle swarm method.
In addition, the estimation unit 602 includes an input layer calculation unit 6021, a hidden layer calculation unit 6022, and an output layer calculation unit 6023, in which:
the input layer calculating unit 6021 is configured to receive the volatilization related data information and directly transmit the received volatilization related data information to the hidden layer calculating unit 6022.
The hidden layer calculation unit 6022 calculates the volatilization time according to the volatilization related data information and the set model parameters, and transmits the calculated volatilization time to the output layer calculation unit 6023.
An output layer calculation unit 6023 configured to output the volatilization time calculated by the hidden layer calculation unit 6022.
That is, the BP neural network is trained prior to use:
the parameter processing unit 605 initializes the parameters of the BP neural network model; the model training processing unit 604 obtains a training set containing multiple sets of volatilization related data information; taking a group of volatilization related data information in the training set as current volatilization related data information, and inputting the current volatilization related data information to an input layer calculation unit 6021 of the BP neural network model; the input layer calculation unit 6021 receives the volatilization related data information and directly transmits the received volatilization related data information to the hidden layer calculation unit 6022; the hidden layer calculation unit 6022 calculates the volatilization time according to the volatilization related data information and the set model parameters, and transmits the calculated volatilization time to the output layer calculation unit 6023; the output layer calculation unit 6023 outputs the volatilization time calculated by the hidden layer calculation unit 6022; the parameter processing unit 605 corrects the parameters of the BP neural network model by using a variable weight particle swarm method according to the output volatilization time serving as a training value.
If the BP neural network model is not converged, the model training processing unit 604 takes the next set of volatilization-related data information in the training set as the current volatilization-related data information, and returns to the step of inputting the current volatilization-related data information to the input layer of the BP neural network model to be repeatedly executed until the BP neural network model is converged.
More specifically, as shown in fig. 8, the internal structure 800 of the parameter processing unit 605 may include: an initialization unit 81 and a correction unit 82. The initialization unit 81 is configured to perform initialization processing on parameters of the BP neural network model. The correcting unit 82 is used for correcting the parameters of the BP neural network model according to the output training values and by using a variable weight particle swarm method.
Further, the correcting unit 82 may include: a particle position setting unit 821, a particle motion calculation unit 822, and a conversion unit 823. Wherein, the particle position setting unit 821 is configured to use the parameter of the BP neural network model as a current position vector of the particle in the variable weight particle swarm method. A particle motion calculation unit 822 for calculating an update velocity of the particle using the current position vector of the particle, the set particle motion velocity, and the variable motion inertia weight; calculating an updated position vector of the particle using the updated velocity of the particle and the current position vector of the particle. A conversion unit 823, configured to convert the updated position vector as a parameter of the modified BP neural network model.
After the BP neural network model is converged, the BP neural network model can be directly used to estimate the volatilization time of indoor harmful substances according to volatilization related data information input by a client, specifically:
the receiving unit 601 obtains volatilization related data information provided by the client; the input layer calculation unit 6021 receives the volatilization related data information and directly transmits the received volatilization related data information to the hidden layer calculation unit 6022; the hidden layer calculation unit 6022 calculates the volatilization time according to the volatilization related data information and the set model parameters, and transmits the calculated volatilization time to the output layer calculation unit 6023; the output layer calculation unit 6023 outputs the volatilization time calculated by the hidden layer calculation unit 6022; the sending unit 603 returns the volatilization time to the client 102 as the estimation result of the volatilization time of the indoor harmful substance.
The embodiment of the present application is a specific description of the server 101, and can be connected to a plurality of clients 102. By applying the scheme of the embodiment of the application, a user does not need to manually estimate or detect the volatilization condition of the indoor harmful substances by using an instrument, and as long as the client 102 sends the current indoor decoration condition, namely the volatilization data information, to the server 101, the server 101 can estimate the volatilization time of the indoor harmful substances and return the indoor harmful substances to the client 102. Therefore, the user of the client 102 can obtain the volatilization time very conveniently and accurately, and the personal safety of the user can be guaranteed to the greatest extent.
As shown in fig. 9, another embodiment of the present application further provides a terminal device 900, which at least includes a processor 901 and a storage medium 902, wherein the processor 901 is configured to execute the steps of the indoor harmful material volatilization time estimation method.
An embodiment of the present application also provides a computer-readable storage medium storing instructions, which when executed by a processor, cause the processor to execute the steps in the indoor harmful material volatilization time estimation method as described above. In practical applications, the computer readable medium may be included in the apparatus/device/system described in the above embodiments, or may exist alone without being assembled into the apparatus/device/system. The above-mentioned computer-readable storage medium carries one or more programs which, when executed, implement the implementation according to the indoor harmful material volatilization time estimation device described with reference to fig. 6 or 7.
According to embodiments disclosed herein, the computer-readable storage medium may be a non-volatile computer-readable storage medium, which may include, for example and without limitation: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing, without limiting the scope of the present disclosure. In the embodiments disclosed herein, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not explicitly recited in the present application. In particular, the features recited in the various embodiments and/or claims of the present application may be combined and/or coupled in various ways, all of which fall within the scope of the present disclosure, without departing from the spirit and teachings of the present application.
The principles and embodiments of the present invention are explained herein using specific examples, which are provided only to help understanding the method and the core idea of the present invention, and are not intended to limit the present application. It will be appreciated by those skilled in the art that changes may be made in this embodiment and its broader aspects and without departing from the principles, spirit and scope of the invention, and that all such modifications, equivalents, improvements and equivalents as may be included within the scope of the invention are intended to be protected by the claims.
Claims (10)
1. A method for estimating volatilization time of indoor harmful substances is characterized by comprising the following steps:
acquiring volatilization related data information provided by a client, wherein the volatilization related data information is data information representing volatilization time of harmful substances influencing the indoor environment;
inputting the acquired volatilization related data information to a reverse relay Broadcasting (BP) neural network model trained in advance, wherein the BP neural network model adopts a variable weight particle swarm method for convergence;
and the BP neural network model calculates volatilization time according to the input volatilization related data information, and returns the volatilization time to the client as an estimation result of the volatilization time of the indoor harmful substances.
2. The method of claim 1, wherein the step of calculating the volatilization time according to the input volatilization-related data information by the BP neural network model comprises:
the input layer of the BP neural network model receives the volatilization related data information and directly transmits the received volatilization related data information to the hidden layer;
the hidden layer calculates the volatilization time according to the volatilization related data information and the set model parameters, and transmits the calculated volatilization time to the output layer;
and the output layer outputs the calculated volatilization time.
3. The method of claim 2, wherein the step of obtaining volatilization related data information is preceded by the step of training a BP neural network model, wherein the step of training the BP neural network model comprises:
acquiring a training set containing multiple groups of volatilization related data information;
initializing the parameters of the BP neural network model;
taking a group of volatilization related data information in the training set as current volatilization related data information, inputting the current volatilization related data information to an input layer of the BP neural network model, and directly transmitting the current volatilization related data information to a hidden layer of the BP neural network model by the input layer;
the hidden layer calculates to obtain a training value according to the current volatilization related data information and the parameters of the BP neural network model, and outputs the calculated training value through an output layer of the BP neural network model;
according to the output training value, correcting the parameters of the BP neural network model by using a variable weight particle swarm method;
and when the BP neural network model is not converged, taking the next group of volatilization related data information in the training set as the current volatilization related data information, and returning the step of inputting the current volatilization related data information to the input layer of the BP neural network model to be repeatedly executed until the BP neural network model is converged.
4. The method according to claim 3, wherein the step of modifying the parameters of the BP neural network model by using a variable weight particle swarm method comprises:
taking the parameters of the BP neural network model as the current position vectors of the particles in the variable weight particle swarm method;
calculating the update speed of the particles by using the current position vector of the particles, the set particle motion speed and the variable motion inertia weight;
calculating an updated position vector of a particle using the updated velocity of the particle and the current position vector of the particle;
and converting the updated position vector to be used as a parameter of the modified BP neural network model.
5. An indoor harmful material volatilization time estimation device, characterized in that the device comprises:
the system comprises a receiving unit, a control unit and a processing unit, wherein the receiving unit is used for acquiring volatilization related data information provided by a client, and the volatilization related data information is data information representing volatilization time of harmful substances influencing indoor;
the estimation unit is used for inputting the acquired volatilization related data information to a reverse relay propagation (BP) neural network model trained in advance, and the BP neural network model adopts a variable weight particle swarm method for convergence; calculating volatilization time according to the input volatilization related data information;
and the sending unit is used for taking the volatilization time as an estimation result of the volatilization time of the indoor harmful substances and returning the estimation result to the client.
6. The apparatus of claim 5, wherein the estimation unit comprises:
the input layer computing unit is used for receiving the volatilization related data information and directly transmitting the received volatilization related data information to the hidden layer computing unit;
the hidden layer calculation unit is used for calculating the volatilization time according to the volatilization related data information and the set model parameters and transmitting the calculated volatilization time to the output layer calculation unit;
and the output layer calculating unit is used for outputting the calculated volatilization time.
7. The apparatus of claim 6, further comprising:
the model training processing unit is used for acquiring a training set containing multiple groups of volatilization related data information; taking a group of volatilization related data information in the training set as current volatilization related data information, and inputting the current volatilization related data information to an input layer of the BP neural network model; when the BP neural network model is not converged, taking the next group of volatilization related data information in the training set as the current volatilization related data information, and returning the step of inputting the current volatilization related data information to the input layer of the BP neural network model to be repeatedly executed until the BP neural network model is converged;
and the parameter processing unit is used for initializing the parameters of the BP neural network model and correcting the parameters of the BP neural network model by using a variable weight particle swarm method according to the output training value.
8. The apparatus of claim 7, wherein the parameter processing unit comprises:
the initialization unit is used for initializing the parameters of the BP neural network model;
and the correcting unit corrects the parameters of the BP neural network model by using a variable weight particle swarm method according to the output training value.
9. The apparatus of claim 8, wherein the correction unit comprises:
the particle position setting unit is used for taking the parameters of the BP neural network model as the current position vectors of the particles in the variable weight particle swarm method;
the particle motion calculation unit is used for calculating the update speed of the particles by using the current position vector of the particles, the set particle motion speed and the variable motion inertia weight; calculating an updated position vector of a particle using the updated velocity of the particle and the current position vector of the particle;
and the conversion unit is used for converting the updated position vector as a parameter of the corrected BP neural network model.
10. A computer-readable storage medium having stored thereon computer instructions, wherein the instructions, when executed by a processor, can implement the steps of the method for estimating volatilization time of indoor harmful material according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911001463.9A CN111047474B (en) | 2019-10-21 | 2019-10-21 | Indoor harmful substance volatilization time estimation method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911001463.9A CN111047474B (en) | 2019-10-21 | 2019-10-21 | Indoor harmful substance volatilization time estimation method, device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111047474A true CN111047474A (en) | 2020-04-21 |
| CN111047474B CN111047474B (en) | 2024-03-22 |
Family
ID=70232327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911001463.9A Active CN111047474B (en) | 2019-10-21 | 2019-10-21 | Indoor harmful substance volatilization time estimation method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111047474B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150328961A1 (en) * | 2013-02-06 | 2015-11-19 | Halla Visteon Climate Control Corp. | Smell generation time predicting device and smell generation time predicting method using the same |
| CN106295802A (en) * | 2016-08-01 | 2017-01-04 | 安徽农业大学 | A kind of Folium Camelliae sinensis based on particle cluster algorithm Optimized BP Neural Network storage method chronological classification |
| CN108921280A (en) * | 2018-04-13 | 2018-11-30 | 广东电网有限责任公司 | A kind of high-voltage circuit-breaker status appraisal procedure based on improvement chaotic particle swarm optimization neural network |
| CN109541143A (en) * | 2018-11-28 | 2019-03-29 | 西安建筑科技大学 | A kind of prediction technique that the constituent element clinker actual constituent transitivity containing volatilization changes over time |
| CN109617888A (en) * | 2018-12-24 | 2019-04-12 | 湖北大学 | A kind of anomalous traffic detection method neural network based and system |
| CN109726889A (en) * | 2018-11-16 | 2019-05-07 | 成都生活家网络科技有限公司 | Appraisal procedure, device and the server of indoor formaldehyde elimination time |
| CN110046834A (en) * | 2019-05-13 | 2019-07-23 | 苏州大学 | Workplace pernicious gas health risk quantitative evaluating system and its appraisal procedure |
| CN110307608A (en) * | 2019-06-11 | 2019-10-08 | 珠海格力电器股份有限公司 | A kind of formaldehyde examination processing method, device, storage medium and air purifier |
-
2019
- 2019-10-21 CN CN201911001463.9A patent/CN111047474B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150328961A1 (en) * | 2013-02-06 | 2015-11-19 | Halla Visteon Climate Control Corp. | Smell generation time predicting device and smell generation time predicting method using the same |
| CN106295802A (en) * | 2016-08-01 | 2017-01-04 | 安徽农业大学 | A kind of Folium Camelliae sinensis based on particle cluster algorithm Optimized BP Neural Network storage method chronological classification |
| CN108921280A (en) * | 2018-04-13 | 2018-11-30 | 广东电网有限责任公司 | A kind of high-voltage circuit-breaker status appraisal procedure based on improvement chaotic particle swarm optimization neural network |
| CN109726889A (en) * | 2018-11-16 | 2019-05-07 | 成都生活家网络科技有限公司 | Appraisal procedure, device and the server of indoor formaldehyde elimination time |
| CN109541143A (en) * | 2018-11-28 | 2019-03-29 | 西安建筑科技大学 | A kind of prediction technique that the constituent element clinker actual constituent transitivity containing volatilization changes over time |
| CN109617888A (en) * | 2018-12-24 | 2019-04-12 | 湖北大学 | A kind of anomalous traffic detection method neural network based and system |
| CN110046834A (en) * | 2019-05-13 | 2019-07-23 | 苏州大学 | Workplace pernicious gas health risk quantitative evaluating system and its appraisal procedure |
| CN110307608A (en) * | 2019-06-11 | 2019-10-08 | 珠海格力电器股份有限公司 | A kind of formaldehyde examination processing method, device, storage medium and air purifier |
Non-Patent Citations (4)
| Title |
|---|
| 堵锡华;王超;: "饮用水中挥发性有机物色谱保留时间的神经网络研究" * |
| 白志鹏,余欢,燕丽,姬亚芹: "室内涂料挥发性污染物的危害和排放规律研究" * |
| 翟春婕;唐松泽;: "基于BP神经网络的木材着火时间预测" * |
| 马驰;: "室内健身场所挥发性污染排放规律动态监测方法研究" * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111047474B (en) | 2024-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017091629A1 (en) | Reinforcement learning using confidence scores | |
| CN110945515A (en) | Intent and slot detection for digital assistants | |
| CN108628904A (en) | A kind of path code, Similar Track search method and device and electronic equipment | |
| Wu et al. | Parametric identification and structure searching for underwater vehicle model using symbolic regression | |
| US20100217734A1 (en) | Method and system for calculating value of website visitor | |
| Grande et al. | Experimental validation of Bayesian nonparametric adaptive control using Gaussian processes | |
| CN108121204A (en) | A kind of self-adaptation control method and system of assembly spacecraft attitude model-free | |
| CN111090688B (en) | Smoothing processing method and device for time sequence data | |
| Baek et al. | Sparse variational deterministic policy gradient for continuous real-time control | |
| JP2012518834A5 (en) | ||
| Grande et al. | Nonparametric adaptive control using Gaussian processes with online hyperparameter estimation | |
| CN114026572A (en) | Error compensation in analog neural networks | |
| CN112990958A (en) | Data processing method, data processing device, storage medium and computer equipment | |
| CN107644433A (en) | Improved closest approach iteration point cloud registration method | |
| CN111047474A (en) | Indoor harmful substance volatilization time estimation method and device and storage medium | |
| CN109857127B (en) | Method and device for calculating training neural network model and aircraft attitude | |
| CN112561351A (en) | Method and device for evaluating task application in satellite system | |
| Mohamed | System identification of flexible aircraft in frequency domain | |
| Vakili et al. | Delayed feedback in kernel bandits | |
| JP2016015586A (en) | LRLS filter | |
| Guzman et al. | Bayesian optimisation for robust model predictive control under model parameter uncertainty | |
| Ananthasayanam et al. | A reference recursive recipe for tuning the statistics of the Kalman filter | |
| CN115730694A (en) | Method and equipment for predicting software workload and storage medium | |
| Cui et al. | The atmospheric model of neural networks based on the improved Levenberg-Marquardt algorithm | |
| Agogino et al. | Efficient credit assignment through evaluation function decomposition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |