Indoor air quality diagnosis and management system using machine learning
The indoor air quality management system uses machine learning and air quality sensors to detect smoking types and classify customers, addressing detection challenges and enabling real-time control in enclosed spaces.
Patent Information
- Application Number
- JP2023551195
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2021-08-10
- Filing Date
- 2021-11-25
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2041-11-25
AI Technical Summary
Existing systems struggle to effectively detect smoking, including both traditional cigarettes and electronic cigarettes, within enclosed spaces like vehicles or accommodation facilities, and fail to provide timely management and control measures when smoking occurs, leading to delayed interventions and poor air quality management.
An indoor air quality diagnosis and management system using air quality sensors and machine learning algorithms to detect the presence and type of smoking, classify customer types, and provide real-time control and notification signals, incorporating models like decision trees, random forests, XGBOOST, and SVMs for accurate smoking detection and air quality analysis.
The system enables immediate detection of smoking types, classifies customer groups for tailored management, and provides real-time control alerts, enhancing air quality monitoring and timely intervention in accommodation facilities.
Smart Images

Figure 0007593691000003 
Figure 0007593691000004 
Figure 0007593691000005
Abstract
Description
[Technical field]
[0001] This specification claims the benefit of priority based on Korean Patent Application No. 10-2021-0024693 filed with the Korean Intellectual Property Office on February 24, 2021, and Korean Patent Application No. 10-2021-0105311 filed with the Korean Intellectual Property Office on August 10, 2021, and all contents disclosed in the documents of said Korean patent applications are incorporated as part of this specification.
[0002] The present invention relates to an indoor air quality diagnosis and management system that measures the indoor air quality in a vehicle or accommodation space, acquires and analyzes air quality data, and diagnoses one or more of the presence or absence of smoking indoors and the type of smoking, thereby detecting the presence or absence of smoking, the type of smoking, and abnormal situations.
[0003] The present invention also relates to a technique that can facilitate management of accommodation facilities by efficiently diagnosing indoor air quality. [Background technology]
[0004] Recently, there has been a trend to encourage people to quit smoking due to health problems such as lung cancer caused by smoking. In particular, smoking in confined indoor spaces such as vehicles and rooms causes great suffering to non-smokers due to secondhand smoke and also has a negative impact on air quality.
[0005] Therefore, despite the strict ban on smoking indoors such as in vehicles and rooms, some smokers still smoke indoors out of sight of others.
[0006] Cigarette smoke spreads over a fairly wide area along with the air flow, so even if someone smokes indoors, it is difficult to immediately detect this, and when others notice and search for the smoker, they often find that the smoker has already finished smoking. In other words, when detecting smoking manually, it is difficult to quickly detect it, so there is a limit to its effectiveness.
[0007] Therefore, there is a need for a system that can immediately detect and notify when a smoker smokes indoors.
[0008] On the other hand, conventional smoking detection systems could only detect cigarettes made before the appearance of electronic cigarettes, and in many cases could not detect electronic cigarettes (including both roll-in cigarettes and liquid cigarettes) because the smoke components are different. Therefore, there is an increasing demand for systems that can detect both cigarettes and electronic cigarettes.
[0009] Furthermore, a characteristic of guest room management at accommodation facilities is that it is difficult for users to quickly intervene and take control measures even if smoking is occurring in the room during the period in which the guest is staying. This results in a problem of delayed detection and action of situations that require room management, such as a deterioration in the condition of the room or the occurrence of an abnormality.
[0010] In addition, in the case of large accommodation facilities, a room management and room control system linked to the usage history of the many customers who visit large accommodation facilities has not yet been applied, so room management cannot be carried out systematically according to the characteristics of each customer, and when multiple situations requiring measures arise, it is difficult for managers to quickly judge and respond to each situation. Summary of the Invention [Problem to be solved by the invention]
[0011] An object of the present invention is to provide an indoor air quality diagnosis and management system that can detect the presence or absence of smoking, the type of smoking, and abnormal situations by measuring the indoor air quality in a vehicle or accommodation space, acquiring and analyzing air quality data, and diagnosing one or more of the presence or absence of smoking indoors and the type of smoking.
[0012] Another object of the present invention is to provide an indoor air quality diagnosis and management system that can analyze management data created for an accommodation facility and air quality data of the accommodation facility to classify the type of customers using the accommodation facility and provide analysis information based on each customer type to an administrator, and that enables real-time control of the guest room status of the accommodation facility and management based on the customer's accommodation usage history. [Means for solving the problem]
[0013] An indoor air quality diagnosis and management system according to one embodiment of the present invention includes an air quality measurement unit that measures indoor air quality in a limited space to obtain air quality data, an air quality analysis unit that analyzes the obtained air quality data based on the results of machine learning of the air quality data depending on the presence or absence of smoking, and an indoor air diagnosis unit that diagnoses one or more of the presence or absence of indoor smoking and the smoking type depending on the analysis results of the air quality data.
[0014] Preferably, the air quality measurement unit may be an air quality sensor (AQS).
[0015] Preferably, the air quality analysis unit may include a detection model generation unit that learns smoking data, which is air quality data when smoking is performed in the room, and non-smoking data, which is air quality data when no smoking is performed in the room, to generate a smoking detection model that can detect the presence or absence of smoking indoors, and that learns cigarette data, which is air quality data when smoking with a cigarette in the room, and cigarette type data, which is air quality data when smoking with an electronic cigarette, to generate a smoking type classification model that can classify smoking types.
[0016] Preferably, the detection model generation unit may generate the smoking detection model or the smoking type classification model using one or more models selected from supervised learning models including a decision tree, a random forest, XGBOOST (Extreme Gradient Boosting), and SVM (Support Vector Machine).
[0017] Preferably, the detection model generation unit may select optimal hyper parameters to be applied to the supervised learning model by one or more of grid search and cross validation.
[0018] Preferably, the indoor air diagnosis unit determines that a smoker is present in the room when smoking is detected a first set number of times within a first set time by the smoking detection model, classifies the smoking type by the smoking type classification model, adds up the number of times classified as a cigarette and the number of times classified as an electronic cigarette, and diagnoses the smoking type based on the number of times classified as a cigarette and the number of times classified as an electronic cigarette that exceeds the majority.
[0019] Preferably, when the indoor air diagnosis unit determines that a smoker is present in the room, after detecting smoking a first set number of times, the indoor air diagnosis unit may not detect smoking for a second set time from the time of the first smoking detection.
[0020] Preferably, the indoor air diagnosis unit can determine that the air quality is "poor" if the ECO2 value measured by the air quality measurement unit is equal to or greater than a first standard value and the TVOC value is equal to or greater than a second standard value, or if the PM10 value is equal to or greater than a third standard value and the PM2.5 value is equal to or greater than a fourth standard value.
[0021] Preferably, the first reference value is 1000 μg / m 3 and the second reference value is 1300 μg / m3 and the third reference value is 80 μg / m 3 and the fourth standard value is 35 μg / m 3 may be also possible.
[0022] Preferably, the indoor air diagnosis unit may determine that an abnormal situation exists when the ECO2 value measured by the air quality measurement unit is equal to or greater than a fifth standard value, the TVOC value is equal to or greater than a sixth standard value, the PM10 value is equal to or greater than a seventh standard value, and the PM2.5 value is equal to or greater than an eighth standard value for a second set number of times within a third set time.
[0023] Preferably, the fifth reference value is 3500 μg / m 3 and the sixth standard value is 4000 μg / m 3 and the seventh standard value is 1700 μg / m 3 and the eighth standard value is 1700 μg / m 3 may be also possible.
[0024] Preferably, the indoor air quality diagnosis and management system may further include an alarm signal output unit that generates and outputs an alarm signal according to an air diagnosis result of the indoor air diagnosis unit.
[0025] An indoor air quality diagnosis method according to one embodiment of the present invention includes an air quality measurement step of measuring indoor air quality in a limited space to obtain air quality data, an air quality analysis step of analyzing the obtained air quality data based on a result of machine learning of the air quality data depending on the presence or absence of smoking, and an indoor air diagnosis step of diagnosing one or more of the presence or absence of indoor smoking and the smoking type depending on the analysis result of the air quality data.
[0026] Preferably, the air quality analysis step may include a detection model generation step of generating a smoking detection model capable of detecting the presence or absence of smoking indoors by learning smoking data, which is air quality data when smoking is performed in the room, and non-smoking data, which is air quality data when no smoking is performed in the room, and generating a smoking type classification model capable of classifying smoking types by learning cigarette data, which is air quality data when smoking with a cigarette in the room, and cigarette type data, which is air quality data when smoking with an electronic cigarette.
[0027] Preferably, the detection model generation step may include a step of generating the smoking detection model or the smoking type classification model using one or more models selected from supervised learning models including a decision tree, a random forest, XGBOOST (Extreme Gradient Boosting), and SVM (Support Vector Machine).
[0028] Preferably, the detection model generation step may further include a step of selecting optimal hyper parameters to be applied to the supervised learning model by one or more of grid search and cross validation.
[0029] Preferably, the indoor air diagnosis step may include the steps of: determining that a smoker is present in the room when smoking is detected a first set number of times within a first set time by the smoking detection model; and classifying the smoking type using the smoking type classification model, adding up the number of times classified as a cigarette and the number of times classified as an e-cigarette, and diagnosing the smoking type based on the number of times classified as a cigarette and the number of times classified as an e-cigarette that exceeds the majority.
[0030] Preferably, the indoor air diagnosis step may further include a step of not detecting smoking for a second set time from the first smoking detection time after a first set number of smoking detections if it is determined that a smoker is present in the room.
[0031] Preferably, the indoor air diagnosis step may include a step of determining that the air quality is "poor" if the ECO2 value measured through the air quality measurement step is equal to or greater than a first standard value and the TVOC value is equal to or greater than a second standard value, or if the PM10 value is equal to or greater than a third standard value and the PM2.5 value is equal to or greater than a fourth standard value.
[0032] Preferably, the first reference value is 1000 μg / m 3 and the second reference value is 1300 μg / m 3 and the third reference value is 80 μg / m 3 and the fourth standard value is 35 μg / m 3 may be also possible.
[0033] Preferably, the indoor air diagnosis step may include a step of determining that an abnormal situation exists when the ECO2 value measured through the air quality measurement step is equal to or greater than a fifth standard value, the TVOC value is equal to or greater than a sixth standard value, the PM10 value is equal to or greater than a seventh standard value, and the PM2.5 value is equal to or greater than an eighth standard value for a second set number of times within a third set time.
[0034] Preferably, the fifth reference value is 3500 μg / m 3 and the sixth standard value is 4000 μg / m 3 and the seventh standard value is 1700 μg / m 3 and the eighth standard value is 1700 μg / m 3 may be also possible.
[0035] Preferably, the indoor air quality diagnosis method may further include a notification signal output step of generating and outputting a notification signal according to an air diagnosis result of the indoor air quality diagnosis step.
[0036] In another embodiment of the present invention, an indoor air quality diagnosis and management system includes an air quality measurement unit that acquires air quality data for each guest room of an accommodation facility; an air quality analysis unit that generates a smoking detection model and a smoking type classification model and analyzes the acquired air quality data based on the results of machine learning of the air quality data; a management data construction unit that constructs management data for the accommodation facility; and a customer type classification unit that classifies the type of customers using the accommodation facility based on the analysis results of the management data and the air quality data. The customer type classification unit can classify the customers into any of a smoking customer group, a non-smoking customer group, and other customer groups, sub-classify customers into the smoking customer group into first smoking customers or second smoking customers, and sub-classify customers into each customer group through logistic regression analysis or cluster analysis.
[0037] Preferably, the management data construction unit can construct the management data based on one or more of the customer's accommodation usage history and the customer's smoking history.
[0038] Preferably, the customer type classifying unit may further classify the customers belonging to the non-smoking customer group into normal customers, caution customers, and risky customers.
[0039] Preferably, the customer type classifying unit can perform the logistic regression analysis or cluster analysis by applying independent variables extracted from the management data.
[0040] Preferably, the customer type classification unit can perform the logistic regression analysis or cluster analysis by setting a criterion for subclassifying the customers as a dependent variable.
[0041] Preferably, the customer type classification unit can perform the logistic regression analysis or cluster analysis to subclassify customers who fall into the smoking customer group and who fall into a range that is more than twice the standard deviation from the average into the second smoking customers, and can perform the logistic regression analysis or cluster analysis to subclassify customers who fall into a range excluding the range that is more than twice the standard deviation from the average into the first smoking customers.
[0042] Preferably, the customer type classification unit performs the logistic regression analysis or cluster analysis to subclassify, among the customers belonging to the non-smoking customer group, those customers who fall into a range that is three or more standard deviations above the average into the risky customers and those customers who fall into a range that is one or more standard deviations above the average into the cautionary customers, and performs the logistic regression analysis or cluster analysis to subclassify, among the customers belonging to the non-smoking customer group, those customers who fall into a range excluding the range that is one or more standard deviations above the average into the normal customers.
[0043] Preferably, the system may further include a control alert transmission unit that transmits a control alert for each guest room of the accommodation facility to the customer based on the classified customer type; and an analysis information providing unit that provides reporting-type analysis information to an administrator of the accommodation facility based on the classified customer type. Effect of the Invention
[0044] According to one aspect of the present invention, indoor air quality in a limited space is measured to obtain air quality data, the obtained air quality data is analyzed based on the results of machine learning of the air quality data based on the presence or absence of smoking, and one or more of the presence or absence of smoking indoors and the type of smoking are diagnosed according to the results of the analysis of the air quality data. Through the measurement of indoor air quality, it is possible to determine not only the presence or absence of smoking indoors, but also the type of smoking, such as cigarettes or e-cigarettes, and further to detect abnormal situations such as fires.
[0045] According to another aspect of the present invention, the abnormality occurrence history for each guest room in the accommodation facility and information regarding the customers and guest rooms included in the associated customer group are provided to the manager via the control service, thereby enabling the manager to easily check the abnormality occurrence history for each guest room and manage the customers and rooms.
[0046] In addition, according to another aspect of the present invention, if abnormal behavior such as smoking in a guest room is detected, a control alarm can be transmitted to customers in real time to prevent unauthorized use of the accommodation facility. [Brief description of the drawings]
[0047] [Figure 1] 1 is a schematic diagram illustrating an indoor air quality diagnosis and management system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram for explaining a Support Vector Machine (SVM) among supervised learning models used in an indoor air quality diagnosis and management system according to an embodiment of the present invention. [Diagram 3] FIG. 11 is a diagram illustrating cross validation, one of the methods for selecting optimal parameters to be applied to a supervised learning model used in an indoor air quality diagnosis and management system according to an embodiment of the present invention. [Figure 4] FIG. 1 is a diagram illustrating a confusion matrix among evaluation indices for evaluating supervised learning models and parameters used in an indoor air quality diagnosis and management system according to an embodiment of the present invention. [Diagram 5] 4 is a diagram illustrating an example of a process of diagnosing indoor air quality in an indoor air diagnosis unit of an indoor air quality diagnosis and management system according to an embodiment of the present invention. FIG. [Figure 6] 2 is a flowchart illustrating a method for diagnosing indoor air quality according to an embodiment of the present invention. [Figure 7] FIG. 2 is a reference diagram for explaining an indoor air quality diagnosis and management system according to another embodiment of the present invention. [Figure 8]1 is a block diagram showing a configuration of a service providing device 10 according to the present invention. [Figure 9] 11 is a reference diagram for explaining a process of classifying types of customers using accommodation facilities according to another embodiment of the present invention. [Figure 10] 11 is a reference diagram for explaining a process of classifying types of customers using accommodation facilities according to another embodiment of the present invention. [Figure 11] 13 is a reference diagram showing a control alarm transmitted to a customer staying at an accommodation facility according to another embodiment of the present invention; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0048] The present invention will be described in detail with reference to the accompanying drawings as follows. Here, repeated descriptions and detailed descriptions of known functions and configurations that may unnecessarily obscure the gist of the present invention will be omitted. The embodiments of the present invention are provided to more completely explain the present invention to those having average knowledge in the art. Therefore, the shapes and sizes of elements in the drawings may be exaggerated for clearer description.
[0049] Throughout the specification, when a part "comprises" a certain element, this does not mean to exclude other elements, but means that it may further include other elements, unless specifically stated to the contrary.
[0050] Furthermore, the term "module" in the specification means a unit that processes one or more functions or operations, which may be embodied in hardware or software, or a combination of hardware and software.
[0051] FIG. 1 is a schematic diagram of an indoor air quality diagnosis and management system according to an embodiment of the present invention.
[0052] Referring to Fig. 1, an indoor air quality diagnosis and management system 100 according to an embodiment of the present invention can be connected to a user's mobile 200 to transmit and receive signals. The indoor air quality diagnosis and management system 100 according to an embodiment of the present invention includes an air quality measurement unit 110, an air quality analysis unit 120, an indoor air diagnosis unit 130, and an alarm signal output unit 140. The indoor air quality diagnosis and management system 100 shown in Fig. 1 is according to an embodiment, and the components shown in Fig. 1 are not limited to the embodiment shown in Fig. 1, and may be added, changed, or deleted as necessary.
[0053] The air quality measurement unit 110 may measure indoor air quality in a limited space to obtain air quality data. In an embodiment, the limited space refers to a space partitioned by walls, doors, etc., such as the inside of a vehicle, a room, etc. The air quality measurement unit 110 may be an air quality sensor (AQS). The air quality sensor is a sensor that measures values of harmful substances contained in the air, such as ECO2, TVOC, PM10, PM2.5, etc. The air quality data generated by measuring the indoor air quality in the air quality measurement unit 110 is used for machine learning in the air quality analysis unit 130 or for analyzing the air quality data.
[0054] The air quality analysis unit 120 may analyze the acquired air quality data based on the results of machine learning of the air quality data depending on whether or not smoking is performed. In one embodiment, the air quality analysis unit 120 may include a detection model generation unit 121 that generates a smoking detection model capable of detecting whether or not smoking is performed indoors by learning smoking data, which is air quality data when smoking is performed indoors, and non-smoking data, which is air quality data when no smoking is performed indoors, and generates a smoking type classification model capable of classifying smoking types by learning cigarette data, which is air quality data when smoking is performed indoors, and cigarette type data, which is air quality data when smoking is performed indoors, and cigarette type data, which is air quality data when smoking is performed with an electronic cigarette.
[0055] The detection model generation unit 121 may analyze air quality data measured by the air quality measurement unit 110 and perform machine learning to generate one or more of a smoking detection model capable of detecting the presence or absence of smoking indoors and a smoking type classification model capable of classifying smoking types. In this case, the detection model generation unit 121 may generate the smoking detection model or the smoking type classification model using one or more models selected from supervised learning models including a decision tree, a random forest, XGBOOST (Extreme Gradient Boosting), and SVM (Support Vector Machine). The models used by the detection model generation unit 121 are as follows.
[0056] Decision Tree A decision tree is a supervised learning model that classifies or predicts all data into small groups according to conditions, and has a tree structure. A decision tree has the advantage that it is easy to explain conditions and branches because it is possible to know through visualization what specific criteria the data was divided by.
[0057] In a tree structure, data that is not divided at all is called a root node, groups that are divided according to conditions are called child nodes, and groups that cannot be divided any further are called final nodes.
[0058] If the final class (the solution to be obtained through the model) is of a discrete type (ex. 0, 1, 2...), it is called a classification tree. If it is of a continuous type (i.e. continuous numbers such as 10.23, 14.56...), it is called a regression tree.
[0059] When the detection model generation unit 121 of the indoor air diagnostic system 100 according to one embodiment of the present invention uses a decision tree among the supervised learning models, the purpose is to classify a single piece of air quality data as being smoking or not, and as being an e-cigarette or a cigarette, so a classification tree is used.
[0060] In the case of classification trees, successive divisions of the data are made using one of several indices, such as the p-value of the chi-squared statistic, the Gini index, or the entropy index of the values of the variables used.
[0061] The smaller the p-value, the greater the heterogeneity between the partitioned nodes, and the smaller the values of the Gini and Entropy indices, the smaller the heterogeneity between the data within a node. The model conditions towards greater heterogeneity between nodes and less heterogeneity between the data within a node.
[0062] There are various options for the criteria for splitting nodes, allowing researchers to choose the direction they wish to go.
[0063] However, decision trees have the disadvantage that the results are difficult to generalize and use because there is a large deviation between models depending on the training data. The random forest technique overcomes this drawback.
[0064] Random Forest Random forest is a supervised learning model based on decision trees, and follows a method of selecting the most frequently appearing result among the results obtained through multiple decision trees (bagging, an ensemble technique).
[0065] In order to overcome the data dependency of existing decision trees, random forests select N pieces of data through sampling with replacement from the training data, and randomly select d variables (e.g. air quality sensor measurements) to be considered for prediction or classification. In this way, prediction and classification are performed through K decision trees. The average or most frequently occurring result for each of the K results is determined as the final value.
[0066] In one embodiment, when random forests and various other machine learning models are generated using Python 3, the results of prediction and classification may vary depending on some hyperparameters (model settings). The model is optimized by selecting the hyperparameters that produce the best results through multiple learning (experiments).
[0067] The supervised learning model usable in the detection model generating unit 121 of the indoor air diagnostic system 100 according to an embodiment of the present invention may consider hyperparameters such as the number of decision trees, the maximum number of variables that can be extracted, the depth of the decision tree, the minimum number of data constituting one node, and the minimum number of data required to constitute a final node, and may be selected through a grid search technique. The following table shows an example of hyperparameters of the supervised learning model usable in the detection model generating unit 121 of the indoor air diagnostic system 100 according to an embodiment of the present invention.
[0068] [Table 1]
[0069] XGBOOST XGBOOST (Extreme Gradient Boosting) is a supervised learning model that applies gradient boosting techniques based on decision trees and has the advantage of being relatively fast and can be implemented in distributed environments.
[0070] Boosting is an ensemble technique that sequentially performs learning and prediction / classification on multiple models and assigns weights to mispredicted data to improve errors. Gradient boosting uses gradient descent to update weights. Gradient descent initializes values by subtracting a value obtained by differentiating a cost function from an initial value, and repeats this process until the differentiated value is minimized.
[0071] Since XGBOOST's performance is likely to change depending on the learning parameters, it is possible to optimize it by conducting experiments on various parameters.
[0072] SVM SVM (Support Vector Machine) is a supervised learning model that defines a baseline for data classification, and the method of drawing the baseline for classification varies depending on the number of attributes. In this case, the baseline is sometimes called the decision boundary.
[0073] FIG. 2 is a diagram for explaining a Support Vector Machine (SVM) among the supervised learning models used in the indoor air quality diagnosis and management system according to one embodiment of the present invention.
[0074] Referring to FIG. 2, for example, as shown in FIG. 2(a), if there are two variable values used for classification (ex. PM2.5, PM10 only), classification can be done using a two-dimensional graph. Also, as shown in FIG. 2(b), if there are three variable values (ex. PM2.5, PM10, TVOC), classification must be done using a three-dimensional graph. In this case, the decision boundary becomes a plane, not a line. In the case of a supervised learning model available in the detection model generation unit 121 of the indoor air diagnostic system 100 according to an embodiment of the present invention, learning can be done using three features, and if visualized, classification should be done using a graph like FIG. 2(b).
[0075] As shown in FIG. 2, the model has the characteristic that as the number of attributes increases, the model becomes more complex, and the classification criterion 'line' (decision boundary) also gradually becomes higher dimensional (this is called a hyperplane).
[0076] SVM basically learns in a direction that maximizes the margin. The margin means the distance between the separated categories, as shown in FIG. 2(c). As shown in FIG. 2(c), the greater the distance between the two categories, the better the separation. The supervised learning model available in the detection model generating unit 121 of the indoor air diagnostic system 100 according to an embodiment of the present invention can also learn with the aim of maximizing the margin, and as a result, the model of the embodiment showed the best accuracy among SVMs.
[0077] Returning to FIG. 1, the detection model generating unit 121 may select optimal hyper parameters to be applied to the supervised learning model by one or more methods of grid search and cross validation. The method used to select the hyper parameters in the detection model generating unit 121 will be described below.
[0078] Grid Search Grid search is one of the methodologies for searching for optimal model hyperparameters. It has the advantage of being quick and efficient. Grid search is a methodology that searches for optimal parameters that can give the best score (based on the evaluation index) by applying all possible combinations of hyperparameter candidates arbitrarily specified by the user.
[0079] Cross-validation Cross-validation is an optimization methodology to prevent overfitting when small datasets are available.
[0080] FIG. 3 is a diagram illustrating cross validation, one of the methods for selecting optimal parameters to be applied to a supervised learning model used in an indoor air quality diagnosis and management system according to an embodiment of the present invention.
[0081] As shown in FIG. 3, the entire data set is divided into a certain number K and evaluated K times. Since the data to be verified is not fixed and evaluation and verification can be performed on all data sets, accuracy can be improved.
[0082] Returning to FIG. 1, the detection model generation unit 121 may evaluate the accuracy of the supervised learning model and parameters used to generate the smoking detection model or the smoking type classification model by using a confusion matrix. The confusion matrix is described below.
[0083] confusion matrix A confusion matrix is a table used to see how well a classification problem has been solved.
[0084] FIG. 4 is a diagram illustrating a confusion matrix, which is one of the evaluation indices for evaluating a supervised learning model and parameters used in an indoor air quality diagnosis and management system according to an embodiment of the present invention.
[0085] In Figure 4, the columns are the results predicted by the classification model, and the rows are the actual values. If the predicted value is correct, it is displayed as TN or TP, and if it is incorrect, it is displayed as FP or FN.
[0086] Accuracy is an index for judging how close predicted data is to actual data, and in the case of a supervised learning model available in the detection model generating unit 121 of the indoor air diagnostic system 100 according to an embodiment of the present invention, the accuracy can be used because the amount of detection is more important than the pressure on false detection. The formula for accuracy is as follows.
[0087]
number
[0088] As described above, the detection model generation unit 121 of the air quality analysis unit 120 generates a smoking detection model that can detect the presence or absence of indoor smoking or a smoking type classification model that can classify smoking types through machine learning, so that the indoor air diagnosis unit 130 can use this to diagnose indoor air quality.
[0089] The indoor air diagnosis unit 130 diagnoses one or more of the presence or absence of smoking indoors and the type of smoking according to the analysis result of the air quality data by the air quality analysis unit 120. In one embodiment, the indoor air diagnosis unit 130 may determine that a smoker is present indoors when smoking is detected a first set number of times within a first set time by a smoking detection model. For example, the indoor air diagnosis unit 130 may detect the presence or absence of smoking by a smoking detection model, but may determine that a smoker is present indoors when smoking is detected three or more times within two minutes. Here, the first set time or the first set number of times may be set to an appropriate value for determining that a smoker is present indoors, and may be set by the user as necessary.
[0090] In one embodiment, the indoor air diagnosis unit 130 may classify the smoking type according to the smoking type classification model, sum up the number of times classified as a cigarette and the number of times classified as an electronic cigarette, and diagnose the smoking type based on the number of times classified as a cigarette and the number of times classified as an electronic cigarette that exceeds the majority. For example, the indoor air diagnosis unit 130 may diagnose the smoking type as a cigarette when the number of times classified as a cigarette is three and the number of times classified as an electronic cigarette is one within two minutes.
[0091] In addition, in one embodiment, the indoor air diagnosis unit 130 may not detect smoking for a second set time from the first smoking detection time after detecting smoking a first set number of times when it determines that a smoker is present in the room. For example, if smoking is detected three or more times within two minutes and it is determined that a smoker is present in the room, it may not detect smoking for six minutes from the first smoking detection time. The reason for not detecting smoking for a certain period of time is to prevent repeated detection of smoking for one smoking act. In one embodiment, the second set time is set to six minutes, which is set in consideration of the average smoking time of four minutes and the normal air quality recovery time of two minutes. However, the second set time is not limited to this and may be set to an appropriate value and may be set by the user as necessary.
[0092] The indoor air diagnostic unit 130 may determine that the air quality is poor if the ECO2 value measured through the air quality measurement unit 110 is equal to or greater than a first reference value and the TVOC value is equal to or greater than a second reference value, or if the PM10 value is equal to or greater than a third reference value and the PM2.5 value is equal to or greater than a fourth reference value. That is, the indoor air diagnostic unit 130 may determine that the air quality is poor and not simply due to smoking if the measured values measured through the air quality measurement unit 110 exceed a predetermined threshold value. For example, the first reference value is 1000 μg / m 3 and the second reference value is 1300 μg / m 3 and the third reference value is 80 μg / m 3 and the fourth standard value is 35 μg / m 3 An example of the third standard value for PM10 is 80 μg / m 3 and 35μg / m, which is an example of the fourth standard value for PM2.5. 3are examples of values selected according to the air quality standards specified by the Korea Meteorological Administration. Meanwhile, in the case of ECO2 and TVOC, since there are no separate grades set by the Korea Meteorological Administration or international standards, examples of the first standard value of ECO2 and the second standard value of TVOC were set based on the values calculated by averaging the observed values of TVOC and ECO2 when PM10 and PM2.5 are at the "poor air quality" level in the air quality observation values collected through experiments. These first to fourth standard values are not limited to the values given as examples, and the first to fourth standard values can be set according to the needs of the user.
[0093] In addition, the indoor air diagnosis unit 130 may determine that an abnormal situation exists when the ECO2 value measured by the air quality measurement unit 110 is equal to or greater than the fifth standard value, the TVOC value is equal to or greater than the sixth standard value, the PM10 value is equal to or greater than the seventh standard value, and the PM2.5 value is equal to or greater than the eighth standard value for a second set number of times within a third set time. Here, the abnormal situation may refer to a situation in which the concentration of indoor harmful substances becomes extremely high due to a fire or the like. For example, the fifth standard value is 3500 μg / m 3 and the sixth standard value is 4000 μg / m 3 and the seventh standard value is 1700 μg / m 3 and the eighth standard value is 1700 μg / m 3 The fifth to eighth reference values are examples of experimental measurements taken when an abnormal situation occurs to the extent that a carbon dioxide alarm actually sounds. The fifth to eighth reference values are not limited to the values given as examples, and the fifth to eighth reference values may be set according to the needs of the user. In this case, the fifth reference value is a value larger than the first reference value, the sixth reference value is a value larger than the second reference value, the seventh reference value is a value larger than the third reference value, and the eighth reference value is a value larger than the fourth reference value. Here, the third set time or the second set number of times can be set to an appropriate value for determining an abnormal situation, and can be set by the user as necessary. For example, the third set time may be two minutes, and the second set number of times may be three times.
[0094] As described above, the process of diagnosing indoor air quality by the indoor air diagnosis unit 130 of the indoor air quality diagnosis and management system 100 according to an embodiment of the present invention has been described through the embodiments. The process of diagnosing indoor air quality by the indoor air diagnosis unit 130 will be described in more detail with reference to FIG. 5 below.
[0095] FIG. 5 is a diagram illustrating an example of a process of diagnosing indoor air quality in an indoor air diagnostic unit of an indoor air quality diagnosis and management system according to an embodiment of the present invention.
[0096] 5, when air quality data measured by the air quality measurement unit 110 is input to the indoor air diagnosis unit 130, it first determines whether the ECO2 value is equal to or greater than the fifth reference value (X1), the TVOC value is equal to or greater than the sixth reference value (Y1), the PM10 value is equal to or greater than the seventh reference value (Z1), and the PM2.5 value is equal to or greater than the eighth reference value (W1) for the second set number of times within the third set time (S501). If the second set number of times within the third set time is reached, it detects an abnormal situation (S502).
[0097] If no abnormal condition is detected, the air quality data is passed through a smoking detection model (S503). If smoking is detected as a result (S504), the coefficient is increased by 1 (S505), and if the coefficient is 3 or greater (S506), the air quality data is entered into a smoking type classification model (S507). This classifies the smoking type as either a cigarette or an e-cigarette (S508).
[0098] If smoking is not detected in step S505, the air quality data is judged as to whether the ECO2 value is equal to or greater than the first standard value X2, the TVOC value is equal to or greater than the second standard value Y2, or the PM10 value is equal to or greater than the third standard value Z2, and the PM2.5 value is equal to or greater than the fourth standard value W2 (S509). If the values are equal to or greater than the first to fourth standard values, the air quality is judged to be "poor" (S510), otherwise the air quality diagnosis process ends.
[0099] Returning to FIG. 1 again, the notification signal output unit 140 generates and outputs a notification signal based on the air diagnosis result of the indoor air diagnosis unit 130. For example, the notification signal output unit 140 can notify the presence of a smoker by issuing an alarm through a speaker or displaying it on a display board. Alternatively, the notification signal output unit 140 can transmit a short message or a push message to the mobile 200 of a manager or user via wireless communication to notify the presence of a smoker. This allows a manager or user in the vicinity to quickly and accurately confirm the presence of a smoker.
[0100] 6 is a flow chart illustrating a method for diagnosing indoor air quality according to an embodiment of the present invention. Referring to FIG. 6, when the method for diagnosing indoor air quality according to an embodiment of the present invention is started, first, indoor air quality in a limited space is measured to obtain air quality data (S610). Then, the obtained air quality data is analyzed based on the results of machine learning of the air quality data according to the presence or absence of smoking (S620).
[0101] Next, depending on the analysis result of the air quality data, the presence or absence of smoking indoors and / or the type of smoking are diagnosed (S630).
[0102] Thereafter, a notification signal is generated and output according to the indoor air diagnosis result of the indoor air diagnosis step (S640).
[0103] The indoor air quality diagnosis method according to one embodiment of the present invention can be realized by each component of the above-mentioned indoor air quality diagnosis and management system, and since the indoor air quality diagnosis method according to one embodiment of the present invention diagnoses indoor air quality in the same manner as the above-mentioned indoor air quality diagnosis and management system, detailed descriptions of the indoor air quality diagnosis method according to one embodiment of the present invention will be omitted to avoid duplication.
[0104] The method for diagnosing indoor air quality according to an embodiment of the present invention may be realized as a program and stored in a computer-readable form on a recording medium such as a CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, Secure Digital (SD) card, micro SD card, or Universal Serial Bus (USB) memory.
[0105] The method for diagnosing indoor air quality according to an embodiment of the present invention may be implemented in the form of a web-based program or an application installed on a mobile terminal. Also, a program implementing the method for diagnosing indoor air quality according to an embodiment of the present invention may be installed in a system for diagnosing and managing indoor air quality according to an embodiment of the present invention.
[0106] FIG. 7 is a reference diagram for explaining an indoor air quality diagnosis and management system (hereinafter, referred to as a management system) according to another embodiment of the present invention.
[0107] 7, the management system is a system applied to a general accommodation facility, and such a management system may include a service providing device 10, an air quality measuring device 20 that may be installed in each guest room of the accommodation facility, and an administrator terminal 30 for transmitting information related to the accommodation facility to a user (or administrator). Such a service providing device 10 may be connected to the air quality measuring device 20 and the administrator terminal 30 via a network.
[0108] The service providing device 10 corresponds to a computing device managed by a service provider that provides an accommodation management service using air quality analysis, and may be realized as a server. Here, the service providing device 10 is connected to the air quality measuring device 20 and the manager terminal 30 via a network, and can provide an accommodation management service that can perform integrated control of each guest room of the accommodation that the manager must manage by analyzing air quality data of the accommodation collected from the connected air quality measuring devices 20-1, 20-2, 20-3, ... and transmitting the analysis result and customer type information according to the analysis result to the manager terminal 30. Here, each of the air quality measuring devices 20-1, 20-2, 20-3, ... may be installed in, for example, guest rooms 101, 102, and 103 of the accommodation. In addition, the manager terminal 30 is a computing device carried, managed, or operated by the accommodation manager. The manager terminal 30 according to the present invention is a computing device that includes a display device and is used as a means for providing an air quality control service to the manager, and may correspond to, for example, an electronic device such as a smartphone, a tablet PC, or a desktop PC. However, such examples are not intended to limit the scope of the present invention, and any computing device capable of providing visual information to an administrator via a display device should be construed as an administrator terminal 30 according to the present invention.
[0109] FIG. 8 is a block diagram showing the configuration of the service providing device 10 according to the present invention.
[0110] 2, the service providing apparatus 10 according to the present invention may include a management data creating unit 210, an air quality measuring unit 220, an air quality analyzing unit 230, and a customer type classifying unit 240. Also, the service providing apparatus 10 according to the present invention may further include a control alarm transmitting unit 250 and an analysis information providing unit 260 according to an embodiment.
[0111] The management data constructing unit 210 may construct management data for accommodation facilities. In one embodiment, the management data constructing unit 210 may construct the management data based on one or more of a customer's accommodation usage history and a smoking history. Here, information included in the constructed management data may be used in a customer type classification process performed according to one embodiment of the present invention.
[0112] The air quality measurement unit 220 may obtain air quality data for each guest room of the accommodation facility. In one embodiment, the air quality measurement unit 220 may collect air quality data from an air quality measurement device 20 installed in each guest room of the accommodation facility where the air quality is to be measured. Here, the data collected from the air quality measurement device 20 may include at least one of a fine dust concentration, a carbon dioxide concentration, and a volatile organic compound concentration. In addition, the air quality measurement device 20 may be configured to include an air quality sensor (AQS). The air quality sensor is a sensor that measures values of harmful substances such as ECO2, TVOC, PM10, and PM2.5 contained in the air, and the air quality data collected from the air quality measurement device 20 may be used for machine learning in the air quality analysis unit 230 or for analyzing the air quality data.
[0113] The air quality analysis unit 230 can analyze the acquired air quality data based on the results of machine learning of the air quality data.
[0114] In one embodiment, the air quality analysis unit 230 can diagnose an abnormal situation for each guest room of the accommodation facility. For example, if the ECO2 value measured by the air quality measuring device 20 is equal to or greater than a first reference value and the TVOC value is equal to or greater than a second reference value, or if the PM10 value is equal to or greater than a third reference value and the PM2.5 value is equal to or greater than a fourth reference value, the air quality of the accommodation facility can be diagnosed as being in an abnormal state. That is, if the measured value measured by the air quality measuring device 20 exceeds a predetermined threshold, the air quality analysis unit 230 can diagnose that the air quality is in an abnormal state, not simply due to smoking. To give a specific example, the first reference value is 1000 μg / m 3 and the second reference value is 1300 μg / m 3 and the third reference value is 80 μg / m 3 and the fourth standard value is 35 μg / m 3 An example of the third standard value for PM10 is 80 μg / m 3 and 35μg / m, which is an example of the fourth standard value for PM2.5. 3 are examples of values selected according to the air quality standards designated by the Japan Meteorological Agency. Meanwhile, in the case of ECO2 and TVOC, since there are no separate grades set by the Japan Meteorological Agency or international standards, examples of the first standard value of ECO2 and the second standard value of TVOC are set based on values calculated by averaging the observed values of TVOC and ECO2 when PM10 and PM2.5 are at the level of abnormal air quality in the air quality observation values collected through experiments. The first to fourth standard values are not limited to the values given as examples, and the first to fourth standard values may be changed as necessary.
[0115] The air quality analysis unit 230 may diagnose that the air quality of the accommodation facility is in a dangerous state when the ECO2 value, TVOC value, PM10 value, PM2.5 value, and PM3.5 value measured by the air quality measuring device 20 are equal to or greater than the fifth standard value, the sixth standard value, the sixth standard value, the seventh standard value, and the eighth standard value for a first set number of times within a first set time. Here, the dangerous state may refer to a situation in which the concentration of indoor harmful substances becomes extremely high due to a fire or the like. As a specific example, the fifth standard value is 3500 μg / m 3 and the sixth standard value is 4000 μg / m 3 and the seventh standard value is 1700 μg / m 3 and the eighth standard value is 1700 μg / m 3 The fifth to eighth reference values are examples of experimentally measured values when the state is dangerous enough that a carbon dioxide alarm actually sounds. The fifth to eighth reference values are not limited to the values given as examples, and the fifth to eighth reference values may be changed as necessary. In this case, the fifth reference value is a value larger than the first reference value, the sixth reference value is a value larger than the second reference value, the seventh reference value is a value larger than the third reference value, and the eighth reference value is a value larger than the fourth reference value. Here, the first set time or the first set number of times can be set to an appropriate value for diagnosing a dangerous state, and may be changed as necessary. For example, the first set time may be two minutes, and the first set number of times may be three times.
[0116] In one embodiment, the air quality analysis unit 230 may include a classification model generation unit that generates a model for classifying an air quality abnormality type based on a result of machine learning on the acquired air quality data. Such a classification model generation unit may generate a smoking detection model or a smoking type classification model using one or more learning models selected from learning models including a decision tree, a random forest, an extreme gradient boosting (XGBOOST), and a support vector machine (SVM).
[0117] In one embodiment, the classification model generating unit may generate a smoking detection model capable of detecting the presence or absence of smoking in a guest room of an accommodation facility, and a smoking type classification model capable of classifying the smoking type in the guest room of the accommodation facility. Specifically, the classification model generating unit may generate a smoking detection model capable of detecting the presence or absence of smoking indoors by learning smoking data, which is air quality data when smoking occurs indoors, and non-smoking data, which is air quality data when smoking does not occur indoors, or may generate a smoking type classification model capable of classifying smoking types by learning cigarette data, which is air quality data when tobacco smoking occurs indoors, and cigarette type data, which is air quality data when electronic cigarette smoking occurs indoors.
[0118] In one embodiment, the classification model generation unit may select optimal hyper parameters to be applied to the learning model using one or more of grid search and cross validation methods.
[0119] In one embodiment, the air quality analysis unit 230 may determine that a customer at the accommodation facility has smoked if smoking is detected a second set number of times within a second set time period by the smoking detection model. For example, the air quality analysis unit 230 may detect the presence or absence of smoking using the smoking detection model, and may determine that a customer at the accommodation facility has smoked if smoking is detected three or more times within two minutes. Here, the second set time period or the second set number of times may be set to an appropriate value for determining whether or not the customer has smoked, and may be changed as necessary.
[0120] In one embodiment, when the air quality analysis unit 230 determines that a guest of the accommodation facility has smoked, it may not detect smoking for a third set time from the first smoking detection time after the second set number of smoking detections. For example, when smoking is detected three or more times within two minutes and it is determined that the guest has smoked, it may not detect smoking for six minutes from the first smoking detection time. The reason for not detecting smoking for a certain period of time is to prevent repeated smoking detection for one smoking act. In the above example, the third set time is set to six minutes, which is set in consideration of the average smoking time of four minutes and the normal air quality recovery time of two minutes. However, the third set time is not limited to this, and may be set to an appropriate value or may be changed as necessary.
[0121] In one embodiment, the air quality analysis unit 230 diagnoses the final smoking type based on the smoking type classified through the smoking type classification model, and may diagnose the smoking type that shows a higher number of times classified as a cigarette or an e-cigarette as the final smoking type.
[0122] The customer type classification unit 240 may classify the type of customer using the accommodation facility based on the analysis result of the management data and the air quality data. Here, the classified customer type may be used to construct management data, and when transmitted to a manager who manages the accommodation facility, the manager may use the data to provide customized services according to each customer type.
[0123] In an embodiment, the customer type classifier 240 may classify customers into a smoking customer group, a non-smoking customer group, and other customer groups. Here, the smoking customer group includes customers who smoke in the accommodation facility, the non-smoking customer group includes customers who use the accommodation facility normally or whose room is diagnosed as having normal air quality, customers whose room air quality is diagnosed as being abnormal, and customers whose room air quality is diagnosed as being at risk due to fire or abnormal behavior, and the other customer group includes customers whose air quality sensor of the air quality measuring device 20 installed in the room does not change and whose measured value is not transmitted to the service providing device 10. The classification of each customer group may basically be performed by reflecting the analysis result of the air quality analyzer 230. Here, the customer type classifier 240 may use logistic regression analysis or cluster analysis to classify the customer type, which will be described in detail later.
[0124] In one embodiment, the customer type classification unit 240 may subclassify customers belonging to the smoking customer group into first smoking customers or second smoking customers. Here, the first smoking customers may include customers who are detected to smoke less than a certain amount compared to the second smoking customers, and customers who smoke more frequently with electronic cigarettes than with cigarettes, and may refer to customers who have an influence below a standard set for a room of an accommodation facility. Also, the second smoking customers may refer to customers who are detected to smoke more than the first smoking customers, for example, customers who are detected to smoke several times, and may refer to customers who have an influence above a standard set for a room of an accommodation facility.
[0125] In one embodiment, the customer type classification unit 240 can subclassify customers belonging to the non-smoking customer group into normal customers, caution customers, and risk customers. Here, the normal customer can mean a customer whose room's air quality is diagnosed as normal and who uses the accommodation facility normally according to the established rules, and the caution customer can mean a customer whose room's air quality is diagnosed as being abnormal. Also, the risk customer can mean a customer whose room's air quality is diagnosed as being dangerous due to a fire or abnormal behavior in the accommodation facility, or other abnormal situations may occur.
[0126] In one embodiment, the customer type classification unit 240 may subclassify customers into each customer group through logistic regression analysis or cluster analysis. For example, the customer type classification unit 240 may perform logistic regression analysis or cluster analysis by applying independent variables extracted from the management data. Here, the independent variables extracted from the management data may include the number of times a customer has used an accommodation facility in the past, the period of use of the accommodation facility, customer location information within the accommodation facility, room type information used by the customer, the customer's smoking history (including smoking frequency), the current customer's smoking frequency, the installation date of the air quality measuring device 20, and software information of the air quality measuring device 20.
[0127] In addition, the customer type classifier 240 may perform logistic regression analysis or cluster analysis by setting a criterion for subclassifying customers as a dependent variable. According to an embodiment of the present invention, the criterion for subclassifying customers may be standard deviation.
[0128] In this regard, FIGS. 9 and 10 are reference diagrams for explaining a process of classifying types of customers using accommodation facilities according to another embodiment of the present invention.
[0129] 9 and 10, in one embodiment, the customer type classification unit 240 may perform the logistic regression analysis or cluster analysis to subclassify customers who belong to a section that is more than twice the standard deviation from the average among the smoking customer group into second smoking customers, and customers who belong to a section excluding the section that is more than twice the standard deviation from the average among the smoking customer group into the first smoking customers. Also, the customer type classification unit 240 may perform the logistic regression analysis or cluster analysis to subclassify customers who belong to a section that is more than three times the standard deviation from the average among the non-smoking customer group into risky customers, customers who belong to a section that is more than one standard deviation from the average into caution customers, and customers who belong to a section excluding the section that is more than one standard deviation from the average among the non-smoking customer group into normal customers.
[0130] The control alert transmission unit 250 can transmit control alerts for each guest room of the accommodation facility to the customers based on the type of customer classified.
[0131] In relation to this, FIG. 11 is a reference diagram showing a control alarm transmitted to a customer staying at an accommodation facility according to another embodiment of the present invention.
[0132] 11, the control alarm transmission unit 250 can transmit a control alarm to a customer through a display device (e.g., a device capable of outputting a control alarm, such as a TV or a PC) installed in each guest room of the accommodation facility, and can also transmit the control alarm through the customer's terminal. The control alarm can suppress unauthorized conduct, such as smoking, in the guest room, and can also notify the customer in the guest room of a dangerous situation, such as a fire or an increase in the concentration of harmful substances, so that the customer in the guest room can quickly respond.
[0133] The analysis information providing unit 260 can provide reporting-type analysis information to the manager of the accommodation facility based on the type of customer classified.
[0134] In one embodiment, the analysis information providing unit 260 may transmit fault reporting analysis information related to a problem occurring in a guest room of a customer group to a manager. For example, if a fire occurs in a guest room, the report analysis information on the fire occurrence may be transmitted to the manager via email or the manager terminal 30, so that the manager can grasp the problem occurring in the accommodation facility in real time and take measures.
[0135] In one embodiment, the analysis information providing unit 260 can provide reporting-type analysis information including the smoking history of the customer to the administrator. Through this, the administrator can monitor the smoking history of each customer and design a customized service to be provided to each customer. For example, the administrator can design an accommodation provision service that applies a penalty to customers with excessive smoking history for using accommodation facilities, or a reward to customers who use accommodation facilities normally more than a certain number of times.
[0136] Although specific embodiments of the present invention have been illustrated and described above, it is obvious to those skilled in the art that the technical concept of the present invention is not limited to the accompanying drawings and the above description, and that various modifications are possible within the scope of the concept of the present invention. It can be said that such modifications fall within the scope of the claims of the present invention as long as they do not violate the spirit of the present invention.
Claims
1. an air quality measuring unit that measures indoor air quality in a limited space and acquires air quality data; an air quality analysis unit that analyzes the acquired air quality data based on the results of machine learning of the air quality data according to the presence or absence of smoking; and an indoor air diagnostic unit for diagnosing one or more of the presence or absence of indoor smoking and the type of smoking according to an analysis result of the air quality data, The air quality analysis unit is a detection model generation unit that learns smoking data, which is air quality data when indoor smoking is performed, and non-smoking data, which is air quality data when indoor smoking is not performed, to generate a smoking detection model that can detect the presence or absence of indoor smoking, and a smoking type classification model that learns cigarette data, which is air quality data when indoor smoking is performed with a cigarette, and cigarette type data, which is air quality data when indoor smoking is performed with an electronic cigarette, to generate a smoking type classification model that can classify the smoking type, The detection model generation unit includes: An indoor air quality diagnosis and management system, characterized in that, in the parameters (hyper parameters) applied to a supervised learning model, optimal parameters are selected by one or more methods of grid search and cross validation.
2. The detection model generation unit includes:
2. The indoor air quality diagnosis and management system according to claim 1, wherein the smoking detection model or the smoking type classification model is generated using one or more models selected from supervised learning models including a decision tree, a random forest, an extreme gradient boosting (XGBOOST), and a support vector machine (SVM).
3. The indoor air diagnosis unit includes: If smoking is detected a first set number of times within a first set time by the smoking detection model, it is determined that a smoker is present in the room; Classifying the smoking type according to the smoking type classification model, summing the number of times classified as cigarettes and the number of times classified as e-cigarettes, and diagnosing the smoking type based on the number of times classified as cigarettes and the number of times classified as e-cigarettes that exceeds a majority; 2. The indoor air quality diagnosis and management system of claim 1, wherein, when it is determined that a smoker is present in the room, after a first set number of smoking detections, smoking is not detected for a second set time from the first smoking detection.
4. The air quality measurement unit includes: The indoor air quality diagnosis and control system according to claim 1, characterized in that it is an air quality sensor (AQS).
5. The indoor air diagnosis unit includes:
2. The indoor air quality diagnosis and management system of claim 1, wherein the air quality is determined to be “poor” when the ECO2 value measured by the air quality measurement unit is equal to or greater than a first standard value and the TVOC value is equal to or greater than a second standard value, or when the PM10 value is equal to or greater than a third standard value and the PM2.5 value is equal to or greater than a fourth standard value.
6. The first reference value is 1000 μg / m 3 and the second reference value is 1300 μg / m 3 and the third reference value is 80 μg / m 3 and the fourth reference value is 35 μg / m 3 The indoor air quality diagnosis and management system according to claim 5 ,
7. The indoor air diagnosis unit includes:
2. The indoor air quality diagnosis and management system of claim 1, wherein when the ECO2 value measured by the air quality measurement unit is equal to or greater than a fifth standard value, the TVOC value is equal to or greater than a sixth standard value, the PM10 value is equal to or greater than a seventh standard value, and the PM2.5 value is equal to or greater than an eighth standard value for a second set number of times within a third set time, it is determined that an abnormal situation exists.
8. The fifth reference value is 3500 μg / m 3 and the sixth reference value is 4000 μg / m 3 and the seventh standard value is 1700 μg / m 3 and the eighth standard value is 1700 μg / m 3 The indoor air quality diagnosis and management system according to claim 7, characterized in that:
9. The indoor air quality diagnosis and management system according to claim 1 , further comprising an alarm signal output unit for generating and outputting an alarm signal according to an air diagnosis result of the indoor air diagnosis unit.
Citation Information
Patent Citations
Fire alarm system
JP2003162778A
Air quality evaluation system, data structure, and program
JP2020024586A
Smoking management apparatus
KR1020180126187A
Intelligent Sensing System For Indoor Air Quality Analytics
US20180073759A1
Environment measurement device, information processing device, information processing system, information processing method, and program
WO2019239855A1
