KR20230103976A - Method for predicting volume of fine particulate matter in indoor space, program and apparatus therefor - Google Patents

Abstract

An apparatus for predicting the volume of ultrafine dust in an indoor space is disclosed. The apparatus selects feature information having a predetermined correlation with fine dust from among a sensing unit that measures one or more particulate matter and environmental factors, and one or more particulate matter and environmental factors measured through the sensing unit, and converts the selected feature information to artificial It may include a control unit that predicts the volume of ultrafine dust in the indoor space after a predetermined time has elapsed by inputting the information to a predictive model learned in advance based on intelligence. By providing this device, since the volume of ultrafine dust in the indoor space can be predicted in real time, user convenience can be improved.

Description

Method for predicting the volume of ultrafine dust, program and device therefor

The present disclosure relates to techniques for predicting contamination conditions. More specifically, the present disclosure relates to a method for predicting the volume of ultrafine dust in an indoor space, a program and an apparatus therefor.

Dust refers to particulate matter that floats or is blown down in the air. There are natural sources of fine dust, but in general, soot generated when fossil fuels such as coal and oil are burned in boilers and power generation facilities, vehicle exhaust gas, It can be said to be artificial, such as flying dust from construction sites, raw materials in the form of powder in factories, powder components in the handling process of subsidiary materials, and smoke from incineration plants. Fine dust generated in this way floats in the air and is deposited on roads, etc., and has a characteristic of re-dispersion.

Such dust is classified into Total Suspended Particles (TSP), which is less than 50 μm, and Particulate Matter (PM), which has a very small particle size, depending on the particle size. Fine dust is divided into fine dust with a diameter of less than 10 μm (hereinafter referred to as PM10) and ultrafine dust with a diameter of less than 2.5 μm (hereinafter referred to as PM2.5).

If PM10 is about 1/5 to 1/7 smaller than the diameter of a human hair (50 to 70㎛), PM2.5 is very small, about 1/20 to 1/30 of a human hair. Since fine dust is so small that it is invisible to the naked eye, it may have a bad effect on health by staying in the air and penetrating the lungs through the respiratory tract or moving into the body along the blood vessels.

The World Health Organization (WHO) has been presenting air quality guidelines for fine dust (PM10, PM2.5, etc.) since 1987, and in 2013, the International Agency for Research on Cancer (IARC) under the World Health Organization ) designated fine dust as a Group 1 carcinogen that has been confirmed to be carcinogenic to humans (Group 1), and interest in it has recently increased due to health risks such as respiratory diseases, cardiovascular diseases, and asthma caused by fine dust.

Therefore, fine dust that poses a risk to health occurs more and more frequently, and weather forecasts also present information on fine dust by expressing PM10 as fine dust and PM2.5 as ultrafine dust. However, while the climate continuously changes, information on fine dust is not presented quickly, and accurate information is not obtained because the measurement location is different from the area where individual people are located.

A lot of research has been conducted on methods and devices that can accurately identify fine dust information and reduce fine dust, and in particular, the risk of fine dust for indoor areas where people, the elderly and children who are vulnerable to fine dust live. need to reduce

Accordingly, a method for predicting the volume (concentration) of ultrafine dust in an indoor space is required.

Publication No. 10-2021-0081600 (Publication date: 2021.07.02.)

An object of the present disclosure is to provide a method for predicting the volume of ultrafine dust in an indoor space.

The problems to be solved by the present disclosure are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

A method for predicting the volume of ultrafine dust in an indoor space performed by a computing device according to an aspect of the present disclosure for achieving the above-described technical problem is a predetermined correlation with ultrafine dust among one or more particulate materials and environmental factors. selecting feature information having a degree; and predicting the volume of ultrafine dust in the indoor space after a predetermined time has elapsed by inputting the selected feature information to a predictive model pre-learned based on artificial intelligence.

The one or more particulate matter has a diameter of less than or equal to 0.1 μm, particulate matter greater than 0.1 μm and less than or equal to 0.3 μm, particulate matter greater than 0.3 μm and less than or equal to 0.5 μm, particulate matter greater than 0.5 μm and less than or equal to 1.0 μm, greater than 1.0 μm and less than or equal to 2.5 μm. It includes particulate matter of less than μm and particulate matter of more than 2.5 μm and less than 5.0 μm, and the environmental factor may include humidity and temperature of the indoor space.

The selecting of the feature information may include calculating a Pearson's correlation coefficient between the at least one particulate matter and the ultrafine dust and between the environmental factor and the ultrafine dust; and selecting feature information having a value of the calculated Pearson's correlation coefficient greater than or equal to a predetermined value.

The predictive model may include a first network and a second network reflecting output values of the first network.

The first network receives the selected specific information and learns to output time-sequential data, and the second network receives the output time-sequential data and predicts the volume of ultrafine dust in the indoor space. can be learned to do.

The second network may be trained to receive the output time-sequential data, detect and classify an anomaly state of the data.

The second network may be learned by reflecting uncertainty of prediction.

According to one aspect of the present disclosure for achieving the above technical problem An apparatus for predicting the volume of ultrafine dust in an indoor space includes a sensing unit for measuring one or more particulate matter and environmental factors; and selecting feature information having a predetermined correlation with fine dust among one or more particulate matter and environmental factors measured through the sensing unit, inputting the selected feature information into a pre-learned prediction model based on artificial intelligence, for a predetermined time A control unit for estimating the volume of ultrafine dust in the indoor space after passage may be included.

The one or more particulate matter has a diameter of less than or equal to 0.1 μm, particulate matter greater than 0.1 μm and less than or equal to 0.3 μm, particulate matter greater than 0.3 μm and less than or equal to 0.5 μm, particulate matter greater than 0.5 μm and less than or equal to 1.0 μm, greater than 1.0 μm and less than or equal to 2.5 μm. It includes particulate matter of less than μm and particulate matter of more than 2.5 μm and less than 5.0 μm, and the environmental factor may include humidity and temperature of the indoor space.

The control unit calculates Pearson correlation coefficients between the one or more particulate matter and the ultrafine dust and between the environmental factor and the ultrafine dust, and selects feature information having a value of the calculated Pearson correlation coefficient equal to or greater than a predetermined value. can be configured.

The predictive model may include a first network and a second network reflecting output values of the first network, wherein the first network receives the selected specific information and learns to output time-series data, The second network may receive the output time-sequential data and learn to predict the volume of ultrafine dust in the indoor space.

The second network may be trained to receive the output time-sequential data, detect and classify an anomaly state of the data, and may be trained to reflect the uncertainty of prediction.

In addition to this, a computer program stored in a computer readable recording medium for execution to implement the present disclosure may be further provided.

In addition to this, a computer readable recording medium recording a computer program for executing a method for implementing the present disclosure may be further provided.

According to the above-described problem solving means of the present disclosure, the volume (concentration) of ultrafine dust in an indoor space can be accurately predicted over time, so that user convenience can be improved.

The effects of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram for schematically explaining an environment for predicting the volume of ultrafine dust according to the present disclosure.
2 is a block diagram showing the configuration of an apparatus for predicting the volume of ultrafine dust according to the present disclosure.
3 is a sequence diagram illustrating a method for predicting the volume of ultrafine dust in an indoor space according to the present disclosure.
4 to 10 are drawings for explaining in detail a method for predicting the volume of ultrafine dust according to the present disclosure.

Like reference numbers designate like elements throughout this disclosure. The present disclosure does not describe all elements of the embodiments, and general content or overlapping content between the embodiments in the technical field to which the present disclosure belongs is omitted. The term 'unit, module, member, or block' used in the specification may be implemented as software or hardware, and according to embodiments, a plurality of 'units, modules, members, or blocks' may be implemented as one component, It is also possible that one 'part, module, member, block' includes a plurality of components.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case of being directly connected but also the case of being indirectly connected, and indirect connection includes being connected through a wireless communication network. do.

In addition, when a certain component is said to "include", this means that it may further include other components without excluding other components unless otherwise stated.

Throughout the specification, when a member is said to be located “on” another member, this includes not only a case where a member is in contact with another member, but also a case where another member exists between the two members.

Terms such as first and second are used to distinguish one component from another, and the components are not limited by the aforementioned terms.

Expressions in the singular number include plural expressions unless the context clearly dictates otherwise.

In each step, the identification code is used for convenience of description, and the identification code does not explain the order of each step, and each step may be performed in a different order from the specified order unless a specific order is clearly described in context. there is.

Hereinafter, the working principle and embodiments of the present disclosure will be described with reference to the accompanying drawings.

In this specification, the 'ultrafine dust volume prediction device according to the present disclosure' includes all various devices capable of providing results to users by performing calculation processing. For example, the device for predicting the volume of ultrafine dust according to the present disclosure may include a computer, a server device, and a portable terminal, or may be in any one form.

Here, the computer may include, for example, a laptop computer, a desktop computer, a laptop computer, a tablet PC, a slate PC, and the like equipped with a web browser.

The server device is a server that processes information by communicating with an external device, and may include an application server, a computing server, a database server, a file server, a game server, a mail server, a proxy server, and a web server.

The portable terminal is, for example, a wireless communication device that ensures portability and mobility, and includes a Personal Communication System (PCS), a Global System for Mobile communications (GSM), a Personal Digital Cellular (PDC), a Personal Handyphone System (PHS), and a PDA. (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), WiBro (Wireless Broadband Internet) terminal, smart phone ) and wearable devices such as watches, rings, bracelets, anklets, necklaces, glasses, contact lenses, or head-mounted-devices (HMDs). can include

1 is a diagram schematically illustrating an environment 1000 for predicting the volume of ultrafine dust according to the present disclosure.

The environment 1000 for predicting the volume of ultrafine dust may be an environment for predicting the volume of ultrafine dust in an indoor space of various places (PL, eg, a building, factory, home, etc.).

Here, ultrafine dust may be particulate matter having a diameter of 2.5 μm or less (eg, greater than 0.1 μm and less than 2.5 μm), but may include particulate matter of various sizes according to embodiments. Ultrafine dust can also be described as PM2.5.

The environment 1000 for predicting the volume of ultrafine dust may include one or more ultrafine dust volume prediction devices 100. The ultrafine dust volume prediction device 100 is in the indoor space of a specific place PL The concentration (volume) of the particulate matter for each size can be measured using various sensors (Se).

The ultrafine dust volume prediction device 100 may collect data of particulate matter and environmental factors (P1), and may select characteristic information from the collected data (P2).

The ultrafine dust volume prediction device 100 may select feature information correlated with generation of ultrafine dust from among the collected data. Characteristic information may be information correlated with generation of ultrafine dust among collected particulate matter and environmental factors.

The ultrafine dust volume prediction device 100 may predict the concentration (volume) of ultrafine dust in the indoor space by inputting the selected feature information to the prediction model (EM) (P3).

2 is a block diagram showing the configuration of the ultrafine dust volume prediction device 100 according to the present disclosure. The ultrafine dust volume prediction device 100 may include a communication unit 110, an input unit 120, a sensing unit 130, a display 140, a memory 150, and a control unit 190.

Among the components, the communication unit 110 may include one or more components enabling communication with an external device, for example, a broadcast receiving module, a wired communication module, a wireless communication module, a short-distance communication module, and location information. At least one of the modules may be included.

The input unit 120 is for inputting image information (or signals), audio information (or signals), data, or information input from a user, and includes at least one of at least one camera, at least one microphone, and a user input unit. can do. Voice data or image data collected by the input unit 120 may be analyzed and processed as a user's control command.

The sensing unit 130 may measure one or more particulate matter and environmental factors, senses at least one of internal information of the device, surrounding environment information surrounding the device, and user information, and generates a sensing signal corresponding thereto. . Based on these sensing signals, the controller 190 may control driving or operation of the device, or perform data processing, functions, or operations related to an application program installed in the device. In an embodiment, the sensing unit 130 may be implemented separately from the present device and may collect sensing data through the communication unit 110 or the input unit 120 .

As described above, the sensing unit 130 may include an environmental sensor (eg, including at least one of a barometer, a hygrometer, a thermometer, a radiation detection sensor, a heat detection sensor, and a gas detection sensor), a chemical sensor (eg, a health sensor). care sensor, biometric sensor, etc.), proximity sensor, illumination sensor, touch sensor, acceleration sensor, magnetic sensor, gravity sensor (G-sensor) ), gyroscope sensor, motion sensor, RGB sensor, infrared sensor, fingerprint sensor, ultrasonic sensor, optical sensor It may include at least one of a sensor (eg, a camera) and a microphone. Meanwhile, the present device may combine and utilize information sensed by at least two or more of these sensors.

The display 140 may implement a touch screen by forming a mutual layer structure or integrally with the touch sensor. Such a touch screen may function as a user input unit providing an input interface between the device and the user, and may provide an output interface between the device and the user.

The display 140 displays (outputs) information processed by the present device. For example, the display 140 may display execution screen information of an application program (for example, an application) running in the present device, or UI (User Interface) and GUI (Graphic User Interface) information according to such execution screen information. can

The memory 150 may store data supporting various functions of the device and programs for operation of the control unit, and may store input/output data (eg, music files, still images, moving images, etc.) , a plurality of application programs (application programs or applications) running in the present device, data for the operation of the present device, and instructions can be stored. At least some of these application programs may be downloaded from an external server through wireless communication.

The memory 150 may be a flash memory type, a hard disk type, a solid state disk type, a silicon disk drive type, or a multimedia card micro type. micro type), card type memory (eg SD or XD memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable EEPROM (EEPROM) It may include a storage medium of at least one type of a programmable read-only memory (PROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, and an optical disk. In addition, the memory is separate from the present device, but may be a database connected by wire or wireless.

The memory 150 may store a predictive model (EM). The predictive model (EM) may be a pre-learned model, and may be stored in the memory 150 even during learning.

The control unit 190 includes a memory for storing data for an algorithm or a program for reproducing the algorithm for controlling the operation of components in the present device, and at least one processor for performing the above-described operation using the data stored in the memory ( not shown) may be implemented. In this case, the memory and the processor may be implemented as separate chips. Alternatively, the memory and the processor may be implemented as a single chip.

In addition, the control unit 190 may control any one or a combination of the components described above in order to implement various embodiments according to the present disclosure described in FIGS. 3 to 10 below on the present device.

At least one component may be added or deleted corresponding to the performance of the components shown in FIG. 2 . In addition, it will be easily understood by those skilled in the art that the mutual positions of the components may be changed corresponding to the performance or structure of the system.

Meanwhile, each component shown in FIG. 2 means software and/or hardware components such as a Field Programmable Gate Array (FPGA) and Application Specific Integrated Circuit (ASIC).

3 is a sequence diagram illustrating a method for predicting the volume of ultrafine dust in an indoor space according to the present disclosure. 4 to 10 are diagrams for explaining a method for predicting the volume of ultrafine dust according to the present disclosure. Specifically, FIG. 4 exemplarily shows sensing data collected for several days in a predetermined indoor space according to the present disclosure. 5 is a diagram for explaining a method for selecting feature information forming a predetermined correlation with ultrafine dust (PM2.5) sensed according to the present disclosure, and FIG. 6 is a diagram according to the present disclosure. 7(a) to 10(b) are diagrams for explaining the performance of the predictive model (EM) according to the present disclosure in comparison with other predictive models. While describing FIG. 3 , reference will be made together where necessary.

The control unit 190 of the ultrafine dust volume prediction device 100 may select feature information having a predetermined correlation with ultrafine dust from among one or more particulate matter and environmental factors (S310).

The controller 190 determines the volume (concentration) of particulate matter (PM0.3, PM0.5, PM1.0, PM2.5, PM5.0, PM10, etc.) having various sizes through the sensing unit 130 in the indoor space. may be measured, and environmental factors such as temperature and humidity may be measured through the sensing unit 130 in an indoor space, but the embodiment is not limited thereto.

In an embodiment, the control unit 190 may collect data of the sensing unit 130 through the communication unit 110, and may collect corresponding data based on MODBUS RTU communication, but the embodiment is not limited thereto.

Referring to FIG. 4 , the controller 190 may collect data of temperature, humidity, and particulate matter of various sizes (PM0.3, PM0.5, PM1.0, PM2.5, PM5.0) of a specific indoor space. can The collected data collected for 8 days (86,400 seconds per day) can be about 700,000 per item, and the temperature in the corresponding indoor space can be collected from 20 to 26 degrees, and the humidity from 10% to 50%. can In addition, the controller 190 may collect volume (concentration) data of the particulate matter, wherein the particulate matter has a diameter of 0.3 μm or less (PM0.3), a diameter greater than 0.3 μm and a diameter of 0.5 μm or less (PM0.5), It may include particulate matter such as greater than 0.5 μm and less than or equal to 1.0 μm (PM1.0), greater than 1.0 μm and less than or equal to 2.5 μm (PM2.5), greater than 2.5 μm and less than or equal to 5.0 μm (PM5.0).

Variables Data Count Mean Standard Deviation Min Max Temperature 691,054 22.877 1.012 20.04 25.94 Humidity 691,054 31.539 8.283 16.7 50.27 PM0.3 691,054 30,028.138 18,870.833 0 65,535.01 PM0.5 691,054 9,000.87 6,227.022 634 46,548.01 PM1.0 691,054 1,512.169 1,094.325 109 7,681 PM2.5 691,054 328.869 256.831 22 1882 PM5.0 691,054 33.253 38.617 0 809

The controller 190 may apply Pearson's correlation coefficient for each item as shown in Equation [1] below in order to check the correlation between ultrafine dust (PM2.5) and collected data.

[Equation 1]

및

는 값의 평균일 수 있다. 즉, 비교되는 두 항목 간의 상관 관계가 피어슨 상관 계수에 의해 결정될 수 있다.where r is the Pearson correlation coefficient, x _i and y _i are the values of the data set,

and

may be the average of the values. That is, a correlation between two items to be compared may be determined by a Pearson correlation coefficient.

Referring to FIG. 5 , collected data lag1 and lag2 may be the same data as ultrafine dust (PM2.5), but shifted by a predetermined time.

The Pearson correlation coefficient was 0.39 between ultrafine particulate matter (PM2.5) and temperature, -0.24 between ultrafine particulate matter (PM2.5) and humidity, and ultrafine particulate matter (PM2.5) and particulate matter (PM0.3). 0.11 between ultrafine dust (PM2.5) and particulate matter (PM0.5), 0.88 between ultrafine dust (PM2.5) and particulate matter (PM1.0), 0.94, and ultrafine dust (PM2. 5) and ultrafine dust (PM2.5) 1, between ultrafine dust (PM2.5) and particulate matter (PM5.0) 0.9, between ultrafine dust (PM2.5) and lag1, 2 1 can be calculated as

The controller 190 calculates Pearson correlation coefficients between one or more particulate matter and ultrafine dust (PM2.5) and between environmental factors (temperature, humidity, etc.) and ultrafine dust (PM2.5), and calculates the Pearson correlation. Feature information having a coefficient value greater than or equal to a predetermined value may be selected.

As an optional embodiment, the controller 190 may select feature information having a Pearson correlation coefficient of 0.5 or more in the case of a positive correlation, and select feature information having a Pearson correlation coefficient of −0.5 or less in the case of a negative correlation. can Accordingly, the controller 190 may select particulate matter such as PM0.5, PM1.0, PM2.5, PM5.0, lag1, and lag2 corresponding to the corresponding correlation coefficient as feature information.

As an optional embodiment, when the types of collected data exceed 15 types, the controller 190 calculates the Pearson correlation coefficient, and then sets the number of feature information corresponding to the calculated Pearson correlation coefficient to a specific number of types (for example, 8). ) can be selected below. For example, even if the Pearson correlation coefficient of 10 or more types out of 15 types is 0.5 or more or -0.5 or less, the control unit 190 sorts in ascending or descending order based on the degree of correlation in positive and negative correlations, and then sorts only up to 8 types. You can select feature information.

Referring to FIG. 6 , the predictive model EM may include a first network N1 and a second network N2 reflecting output values of the first network N1.

The first network (N1) can be learned to output time-series data by receiving selected specific information, and reflect the characteristics of time-series data and previous data. It is a long short term memory (LSTM)-based network. can

The first network N1 may sequentially receive input data (X ₁ to X _n ) and output time-sequential data (X _{BNN 1} to X _{BNN n} ). The first network N1 may check how much the concentration of ultrafine dust in the indoor space changes from the default time point by using information on whether a predetermined time has elapsed since the time when the ultrafine dust has a default concentration. Here, the default concentration (volume) can be set differently depending on the characteristics of the place. For example, in the case of a clean room, it can be said that the ultrafine dust is below a predetermined value. may be set differently.

The second network N2 may receive time-sequential data output from the first network N1 and learn to predict the volume of ultrafine dust in the indoor space. The second network N2 may be learned by reflecting uncertainty of prediction and may be implemented as a Bayesian neural network (BNN), but the embodiment is not limited thereto.

The second network N2 may perform an operation by receiving output data (X _{BNN 1} to X _{BNN n} ) of the first network N1 . The second network N2 may include an input layer (i), a hidden layer (j), and an output layer (k).

Here, weights and biases may be used for calculation for each connection between nodes between the input layer (i) and the hidden layer (j) and between the hidden layer (j) and the output layer (k).

(가령, UC1, UC2)이고, 바이어스가

이며, 히든 레이어(j) 및 출력 레이어(k) 사이의 가중치가

(가령, UC3, UC4)이고, 바이어스를

라고 할 수 있다.Referring to Equation 2 below, the weight between the input layer (i) and the hidden layer (j) is

(e.g., UC1, UC2), and the bias is

, and the weight between the hidden layer (j) and the output layer (k) is

(e.g., UC3, UC4), and the bias

can be said

[Equation 2]

는 입력레이어(i) 및 히든 레이어(j) 사이의 가중치의 총합이고,

는 히든 레이어(j) 및 출력레이어(k) 사이의 가중치의 총합일 수 있다. 각 변수(

,

,

,

)는 정규분포를 따르고 정규분포에서 샘플링될 수 있으며, 가중치(

,

)는 정규분포에서 샘플링될 수 있다.

is the sum of the weights between the input layer (i) and the hidden layer (j),

may be the sum of weights between the hidden layer (j) and the output layer (k). Each variable (

,

,

,

) follows a normal distribution and can be sampled from a normal distribution, and the weights (

,

) can be sampled from a normal distribution.

Each weight can be calculated based on a stochastic distribution by [Equation 3] below.

[Equation 3]

Since the control unit 190 can calculate each weight based on a probabilistic distribution, it can learn the prediction model by reflecting the uncertainty of prediction. The control unit 190 may use Equation 3 to finally determine a stochastic anomaly and/or a not anomaly. That is, the control unit 190 learns the first network including label information, receives the learned output value, and based on the corresponding output value, the controller 190 uses a Bayesian Neural Network (BNN) to determine an anomaly and/or a non-anomaly. can decide

The control unit 190 may use label data corresponding to the collected data at a predetermined time to detect whether the data at a predetermined time is in an abnormal state or non-abnormal state compared to the label data, and classify the data.

) 및 우도 함수(

,likelihood function)는 아래 [식 4]에 따라 산출될 수 있다. 사전 확률(

)은 새로운 데이터(D) 를 관찰하기 전에 미리 인식하고 있는 사전 분포일 수 있으며,

는 각 변수(

,

,

,

)를 나타내고, L은 제2 네트워크(N2)의 파라미터 개수를 나타내고, N 은 학습 데이터수를 나타내고, y_n은 nth 데이터의 출력값, f(x)_n 은 nth 데이터의 모델 출력값을 나타낼 수 있다.In [Equation 3], the prior probability (

) and the likelihood function (

,likelihood function) can be calculated according to [Equation 4] below. prior probability (

) may be a prior distribution that is recognized in advance before observing new data (D),

is each variable (

,

,

,

), L represents the number of parameters of the second network N2, N represents the number of training data, y _n represents the output value of nth data, and f(x) _n represents the model output value of nth data.

[Equation 4]

,

,

는 적분식이고, 초기 입력값이 제1 네트워크(N1)를 통해 출력되는 값이므로, 연산이 어려울 수 있으므로, VI(variational inference) 기반으로 적분식의 값을 근사화하는 방법으로 산출할 수 있다.also,

Is an integral equation, and since the initial input value is a value output through the first network N1, it may be difficult to calculate, so it can be calculated by approximating the value of the integral equation based on VI (variational inference).

는 아래 [식 6]으로 대체될 수 있다.Therefore,

can be replaced by [Equation 6] below.

[Equation 6]

,

,

는 근사화 연산을 위해 필요한 파라미터일 수 있다.Here, Q may be a probability function for approximation. Q may follow a probability distribution such as an accessible normal distribution. F is the result value,

may be a parameter required for an approximation operation.

7(a) and 7(b) show result graphs showing ultrafine dust volume prediction after 1 hour in a specific indoor space compared with other algorithms using the predictive model (EM) according to the present disclosure.

Referring to the root mean square error (RMSE) of the predictive model (EM) according to the present disclosure, it can be said that the predictive accuracy of the predictive model (EM) according to the present disclosure is the best compared to other algorithms.

8(a) and 8(b) show result graphs showing ultrafine dust volume prediction after 2 hours in a specific indoor space compared with other algorithms using the prediction model (EM) according to the present disclosure, and FIG. 9 Is a table explaining the performance of the predictive model (EM) according to the present disclosure compared to other algorithms.

For model evaluation, the controller 190 may use mean square error (MSE), RMSE, mean absolute error (MAE), mean absolute percentage error (MAPE), coefficient of determination (R2), and the like. The prediction accuracy of the predictive model (EM) according to the present disclosure is the highest, and it can be confirmed that the accuracy is also the highest.

10(a) and 10(b) show result graphs showing ultrafine dust volume prediction after 3 hours in a specific indoor space compared with other algorithms using the prediction model (EM) according to the present disclosure.

7(a) to 10(b), the volume of ultrafine dust in a specific indoor space can be measured in real time by the predictive model (EM) according to the present disclosure, so that a plan to prepare for indoor environmental pollution is made. It can be set, and user convenience can be improved.

When the prediction model (EM) according to the present disclosure is applied, it can be confirmed that the prediction performance and accuracy of ultrafine dust are the highest than when other algorithms are applied.

Meanwhile, the disclosed embodiments may be implemented in the form of a recording medium storing instructions executable by a computer. Instructions may be stored in the form of program codes, and when executed by a processor, create program modules to perform operations of the disclosed embodiments. The recording medium may be implemented as a computer-readable recording medium.

Computer-readable recording media include all types of recording media in which instructions that can be decoded by a computer are stored. For example, there may be read only memory (ROM), random access memory (RAM), magnetic tape, magnetic disk, flash memory, optical data storage device, and the like.

As above, the disclosed embodiments have been described with reference to the accompanying drawings. Those skilled in the art to which the present disclosure pertains will understand that the present disclosure may be implemented in a form different from the disclosed embodiments without changing the technical spirit or essential features of the present disclosure. The disclosed embodiments are illustrative and should not be construed as limiting.

100: Ultrafine dust volume prediction device,
190: control unit.

Claims (11)

As a method of predicting the volume of ultrafine dust in an indoor space performed by a computing device,
selecting characteristic information having a predetermined correlation with ultrafine dust from among one or more particulate matter and environmental factors; and
Predicting the volume of ultrafine dust in the indoor space after a predetermined period of time by inputting the selected feature information into a prediction model pre-learned based on artificial intelligence. According to claim 1,
The one or more particulate matter has a diameter of less than or equal to 0.1 μm, particulate matter greater than 0.1 μm and less than or equal to 0.3 μm, particulate matter greater than 0.3 μm and less than or equal to 0.5 μm, particulate matter greater than 0.5 μm and less than or equal to 1.0 μm, greater than 1.0 μm and less than or equal to 2.5 μm. particulate matter of less than μm and particulate matter of greater than 2.5 μm and less than or equal to 5.0 μm, wherein the environmental factors include humidity and temperature of the indoor space;
The step of selecting the feature information,
calculating Pearson's correlation coefficients between the one or more particulate matter and the ultrafine dust and between the environmental factor and the ultrafine dust; and
A method for predicting the volume of ultrafine dust comprising the step of selecting feature information having a value of the calculated Pearson's correlation coefficient equal to or greater than a predetermined value. According to claim 2,
The predictive model,
A first network and a second network reflecting output values of the first network,
The first network receives the selected specific information and learns to output time-series data,
The second network receives the outputted time-sequential data and learns to predict the volume of ultrafine dust in the indoor space. According to claim 3,
The second network,
A method for predicting the volume of ultrafine dust that receives the output time-series data and learns to detect and classify an anomaly state of the data. According to claim 4,
The second network,
A method for predicting the volume of ultrafine dust that is learned by reflecting the uncertainty of prediction Combined with a computer, which is hardware, and stored in a medium to execute the method of any one of claims 1 to 5, ultrafine dust volume prediction program. As a device for predicting the volume of ultrafine dust in an indoor space,
a sensing unit for measuring one or more particulate matter and environmental factors; and
Characteristic information having a predetermined correlation with fine dust is selected from among one or more particulate matter and environmental factors measured through the sensing unit, and the selected characteristic information is input into a pre-learned prediction model based on artificial intelligence, and a predetermined time elapses. Ultrafine dust volume prediction device comprising a control unit for predicting the volume of ultrafine dust in the indoor space after. According to claim 7,
The one or more particulate matter has a diameter of less than or equal to 0.1 μm, particulate matter greater than 0.1 μm and less than or equal to 0.3 μm, particulate matter greater than 0.3 μm and less than or equal to 0.5 μm, particulate matter greater than 0.5 μm and less than or equal to 1.0 μm, greater than 1.0 μm and less than or equal to 2.5 μm. particulate matter of less than μm and particulate matter of greater than 2.5 μm and less than or equal to 5.0 μm, wherein the environmental factors include humidity and temperature of the indoor space;
The control unit,
A second configured to calculate a Pearson correlation coefficient between the one or more particulate matter and the ultrafine dust and between the environmental factor and the ultrafine dust, and select feature information for which the value of the calculated Pearson correlation coefficient is equal to or greater than a predetermined value. Fine dust volume prediction device. According to claim 8,
The predictive model,
A first network and a second network reflecting output values of the first network,
The first network receives the selected specific information and learns to output time-series data,
The second network receives the outputted time-sequential data and learns to predict the volume of ultrafine dust in the indoor space. According to claim 9,
The second network,
An ultrafine dust volume prediction device that receives the output time-series data and learns to detect and classify an anomaly state of the data. According to claim 10,
The second network,
A device for predicting the volume of ultrafine dust that is learned by reflecting the uncertainty of the prediction.

Images