JP5940018B2 - Data name extraction device and program - Google Patents

Description

  The present invention relates to a data name extraction apparatus and a program for extracting specific data name information from a plurality of data name information.

  Various buildings such as lighting and air conditioners are installed in buildings such as buildings and plants, and data related to these devices may be acquired, and equipment diagnosis and analysis for energy saving may be performed based on the data. . For example, the set temperature, the actually measured temperature, the air conditioning state, the voltage value, the current value, the pressure value, and the like correspond to the data.

  Data related to the device is given a data name for identifying the data. For example, terms related to the location where the device is installed, terms related to the type of data, terms related to the type of device, and the like may be used as the data name. Such a data name may be arbitrarily created by a user of the building, and is not necessarily created by an analyst who performs device diagnosis or analysis. Therefore, even if the data is related to the same type of device or the same type of data, the data name may be created differently for each building. For example, even if it is data indicating temperature, “temperature” may be added to the data name, and “Temp” or the like may be added to the data name. In this way, under the situation where the data name varies depending on the building, the analyst refers to the data name given to the data, selects the name of the data to be analyzed while considering the meaning of the data name, and selects the selected data It is necessary to extract data with names and perform analysis and diagnosis.

  As an apparatus for automatically classifying data names, Patent Literature 1 determines the types of terms constituting the data names by using a dictionary, and extracts only the terms having high importance using the determined types of terms, An apparatus for classifying similar data names based on extracted high-priority terms is disclosed.

JP-A-6-231188

  As described above, even if the data is related to the same type of equipment or the same type of data, different data names may be created. Therefore, the analyst must select the name of the data to be analyzed while taking the meaning of the data name into account. There was a need to choose. Conventionally, this selection operation has to be performed manually by an analyst, and the number of data names (data items) can be several thousand to several tens of thousands. It is not easy to select and categorize the data names, and it places an excessive burden on the analyst. Also, it takes a lot of time to extract data to be analyzed.

  In addition, the names of data related to equipment vary greatly depending on the type of data such as temperature and pressure, the type of equipment, the location of installation, etc. Are not necessarily similar. Therefore, even if the data names are classified based on terms having high importance as in the invention described in Patent Document 1, the data names given to the data to be analyzed cannot always be accurately classified, There may be many data classification errors.

  An object of the present invention is to provide a data name extraction device and a program capable of accurately extracting a data name given to data to be analyzed from a plurality of data names.

  The invention according to claim 1 is configured to include device type terms related to the type of device to be detected by the sensor, signal type terms related to the type of sensor output value, and installation location terms related to the location where the device is installed, Data name information associated with each output value of multiple sensors is acquired, and the same device type category is assigned to device type terms that are synonymous with each other among the device type terms of each data name information, and each data The same signal type category is assigned to the signal type terms that are synonymous with each other among the signal type terms in the name information, and the same installation location category as the installation location terms that are synonymous with each other among the installation location terms in each data name information Category assigning means for assigning output values, and device type categories and signal type categories for making the output values of multiple sensors the same analysis target Data for which analysis target information that defines the combination of installation location categories is acquired, and the same category as the category defined by the analysis target information is assigned from a plurality of data name information to which categories are assigned by the category assignment means And a data name extracting device characterized by comprising extraction means for extracting name information.

  The invention according to claim 2 is the data name extraction device according to claim 1, wherein the analysis target information defines a plurality of combinations in which the device type category and the installation location category are common to each other and the signal type categories are different from each other. The extraction unit extracts a plurality of pieces of data name information in which the device type category and the installation location category are common to each other and the signal type categories are different from each other, and among the extracted plurality of data name information, the place indicated by the installation location term is The same data name information is classified into the same set.

  The invention according to claim 3 is the data name extraction device according to claim 1 or 2, wherein the extraction means includes the device type category, the signal type category, and the installation location category defined by the analysis target information. If data name information with the same category as all categories does not exist in multiple data name information, some of the device category, signal category, and installation location category defined by the analysis target information The data name information to which the same category as the category is assigned is extracted from a plurality of data name information.

  The invention according to claim 4 is the data name extraction device according to any one of claims 1 to 3, wherein the extraction unit specifies the type of the output value based on the unit of the output value of the sensor. The data name information is extracted based on the signal type category corresponding to the specified type.

  The invention according to claim 5 is the data name extraction device according to any one of claims 1 to 3, wherein the extraction means specifies the type of the output value based on the output value of the sensor, and specifies The data name information is extracted based on the signal type category corresponding to the selected type.

  The invention according to claim 6 includes, in the computer, device type terms relating to the type of device to be detected by the sensor, signal type terms relating to the type of sensor output value, and installation location terms relating to the location where the device is installed. Data name information associated with each output value of a plurality of sensors is acquired, and the same device type category is assigned to device type terms that are synonymous with each other among the device type terms of each data name information. The same signal type category is given to the signal type terms that are synonymous with each other among the signal type terms of each data name information, and the same installation location terms that are synonymous with each other among the installation location information of each data name information The step of assigning the installation location category and the device type category and signal type for making the output values of multiple sensors the same analysis target The analysis target information defining the combination of the category and the installation location category is acquired, and the data name information to which the same category as the category defined by the analysis target information is assigned from the plurality of data name information to which the category is assigned. And a step of extracting the program.

  According to the present invention, by assigning a category to the data name information and using the analysis target information in which the combination of the names of the data to be analyzed is defined by the category, the data name given to the data to be analyzed can be changed. It becomes possible to extract accurately. According to the present invention, it is possible to classify data names that cannot be classified only by the similarity of terms.

It is a block diagram which shows an example of the data analysis system which concerns on embodiment of this invention. It is a block diagram which shows an example of the data name extraction apparatus which concerns on embodiment of this invention. It is a figure which shows an example of a data name list. It is a figure which shows an example of the divided | segmented data name list. It is a figure which shows an example of a category dictionary. It is a figure which shows an example of the data name list with a category. It is a figure which shows an example of a data name group pattern. It is a figure which shows an example of the extracted data name set. It is a figure which shows an example of the data name list which concerns on the modification 1. It is a figure which shows an example of the extracted data name set which concerns on the modification 1. FIG. It is a figure which shows an example of the unit and value of data.

[Overall configuration of data analysis system]
First, the overall configuration of a data analysis system in which a data name extraction device according to an embodiment of the present invention is used will be described. FIG. 1 shows an example of a data analysis system. The data analysis system according to the present embodiment acquires data related to various devices installed in the building 200, extracts data to be analyzed based on the data names given to the data, and analyzes based on the extracted data And diagnoses. Although only one building 200 is illustrated in FIG. 1, data may be acquired from a plurality of buildings 200 for analysis and diagnosis.

  The building 200 is, for example, an office building or a plant. In the building 200, various devices such as an air conditioner 202, an illumination 204, and a power receiving / transforming facility 206 are installed. Various devices that detect various data are installed in the devices installed in the building 200, and the data detected by the sensors is transmitted to the data collection device 210 via a communication path N such as a network. The A data name for identifying the data is associated with the data. This data name is created by the user of the building 200, for example.

  Various data collected by the data collection device 210 is stored in the storage device 222 of the data management system 220. Data names associated with the data are listed and stored in the data management system 220 as the data name list 100.

  The data name extraction apparatus 10 extracts a set of data names (analysis target data name set 140) assigned to data to be analyzed from the data name list 100 stored in the data management system 220.

  The data extraction device 230 extracts the data with the data name extracted by the data name extraction device 10 from the storage device 222 of the data management system 220 as the analysis target data 240. Based on the analysis target data 240, the analysis device 250 performs diagnosis and analysis of the analysis target device.

[Data Name Extraction Device 10]
Next, the data name extraction device 10 according to the present embodiment will be described with reference to FIG. The data name extraction device 10 includes a morphological analysis unit 12, a category assignment unit 14, a category storage unit 16, and an extraction unit 18.

  The morphological analysis unit 12 acquires the data name list 100 from the data management system 220, divides each data name included in the data name list 100 into terms, and a divided data name list that is a list of divided data names 110 is created. By using a conventional technique related to natural language processing (for example, morphological analysis software such as MeCab), it is possible to perform segmentation processing.

  FIG. 3 shows an example of the data name list 100. The data name list 100 includes, for example, data names (1) to (19). These data names may be part of the data name list 100 or all of them. For example, the data name list 100 may include thousands to tens of thousands of data names. The numbers (1) to (19) are shown for convenience of explanation, and the numbers are not assigned to the respective data names. Each data name includes, for example, device type terms related to the type of device to be detected by the sensor, signal type terms related to the type of sensor output signal, and installation location terms related to the location where the device is installed. . Examples of device type terms include “air conditioner” and “lighting”, and signal type terms include physical quantities such as “temperature”, “humidity”, and “pressure”, and installation location terms. For example, “B1F”, “2F”, [first floor], and the like are applicable. For example, the data name (1) is “B1F system air conditioner SA temperature measurement”, and the data name is a sensor that measures the temperature of the air supply (SA) of the air conditioner installed in B1F (first floor underground). Indicates the name of the data output from. Since these data names may be created by the user of the building 200, the rules for creating the data names may be different for each building 200. Therefore, even if the data has the same meaning, different terms may be used for the data name or the order of the terms may be different. For example, when different terms are used, it is considered that the data name “B1F system air conditioner SA temperature measurement” in (1) is created as “basement 1st floor system air conditioner SA temperature measurement”. As a case where the order is different, a case where “SA temperature measurement B1F system air conditioner” is created may be considered.

  FIG. 4 shows an example of the divided data name list 110. For example, the data name (1) is divided by each term such as “B1F / system air conditioner / SA / temperature / measurement”. Similarly, other data names are divided by terms.

  The category assigning unit 14 refers to the category dictionary stored in the category storage unit 16 so that each term included in the divided data name includes a category related to the signal type (signal type category) and a category related to the device type (device type). (Category category) or a category related to the installation location (installation location category), and a data name list 120 with category is created. For example, the category assigning unit 14 assigns the same signal type category to the signal type terms that are synonymous with each other among the signal type terms of each data name, and the devices that are synonymous with each other among the device type terms of each data name The same device type category is assigned to the type term, and the same installation location category is assigned to the installation location terms that are synonymous with each other among the installation location terms of each data name.

  FIG. 5 shows an example of the category dictionary. The category dictionary is information that defines the correspondence between standard terms that are expected to be used for data names and categories, and is information that defines synonyms for the standard terms. As described above, since the data name includes the signal type term, the device type term and the installation location term, the standard term is classified into the signal type term, the device type term and the installation location term. A signal type category, a device type category, and an installation location category are defined. Each category is associated with a standard term. In addition, the terms included in the data names include main words with a relatively high importance and modifiers with a relatively low importance, so each category has a category for the main word and a category for the modifier. It is divided into and.

  For example, “temperature”, “humidity”, “control output”, “supply air”, “return air”, “setting”, and “measurement” are defined as standard terms of the signal type. The standard terms “temperature”, “humidity”, and “control output” indicate physical quantities of data and are terms related to signal types, and are therefore classified as signal type terms. In addition, the standard terms “temperature”, “humidity”, and “control output” are terms that directly indicate the physical quantity of data, and thus are defined as major terms having a relatively high importance. It is categorized in the signal main word (sender) category.

  The standard terms “supply air”, “return air”, “setting”, and “measurement” are not terms that directly indicate the physical quantity of the data, but are terms related to the type of signal, and therefore are classified as signal type terms. . In addition, these terms are not terms that directly indicate the physical quantity of data, so they are defined as modifiers with relatively low importance, and are classified into the signal modifier (Shinshu) category of the signal category. Has been. In the example illustrated in FIG. 5, the term “SA” is defined as a synonym for the term “supply air”, and the term “RA” is defined as a synonym for the term “return air”.

  For example, “air conditioner”, “state”, “valve opening”, and the like are defined as standard terms for the device type. Since the standard term “air conditioner” indicates the type of device, it is classified as a device type term, and is a term that directly indicates the type of device. It is defined and classified into the device main word (machine owner) category in the device type category.

  The standard terms "state" and "valve opening" are not terms that directly indicate the type of device, but are terms related to the type of device, so they are classified as device type terms and are relatively low in importance. It is defined as a modifier and is classified into the device modifier (machine repair) category of the device type categories. In the example shown in FIG. 5, the term “OAMD” is defined as a synonym for the term “state”, and the term “valve” is defined as a synonym for the term “valve opening”.

  As standard terms of the installation location, terms related to the installation location such as standard terms “B1F” and “2F” are defined. Since these terms are terms indicating the place where the device is installed, they are defined as main words having a relatively high importance, and are classified into the place main word (placeholder) category. In addition, as a synonym of an installation place, the term which shows the number of floors, such as "1st floor", "1st basement floor", etc. can be considered, for example.

  FIG. 6 shows an example of the data name list 120 with categories. The assigned category will be described by taking the data name “B1F system air conditioner SA temperature measurement” of (1) as an example. Since the term “B1F” is a place main word, the “place owner” category is assigned. Further, the term “B1F” is given the standard term “B1F”, and as a result, the “B1F: owner” category including the standard term is given. Since the term “system air conditioner” is the main device language, the “machine owner” category is assigned. The term “system air conditioner” is given the standard term “air conditioner”, and as a result, the category “air conditioner: machine owner” including the standard term is given. Since the term “SA” is a signal modifier, the “Shinshu” category is assigned. In addition, the term “SA” is assigned the standard term “air supply”, and as a result, the “air supply: credit” category including the standard term is assigned. Since the term “temperature” is the main signal word, the “follower” category is assigned. Further, the term “temperature” is given the standard term “temperature”, and as a result, the “temperature: believer” category including the standard term is given. Since the term “measurement” is a signal modifier, the “Shinshu” category is assigned. Further, the term “measurement” is given the standard term “measurement”, and as a result, the category “measurement: Shinshu” containing the standard term is given. Other data name terms also have a category including standard terms.

  The extraction unit 18 refers to the data name set pattern 130 to extract a set of data names given to the same data to be analyzed from the category-attached data name list 120, and the data included in the set An analysis target data name set 140 indicating the name is output. The data name combination pattern 130 is information that defines a combination of a device type category, a signal type category, and an installation location category assigned to data names of data to be analyzed. For example, the extraction unit 18 extracts a data name to which the same category as the category defined by the data name set pattern 130 is assigned from the data name list with category 120. The data name set pattern 130 corresponds to an example of analysis target information.

  FIG. 7 shows an example of the data name set pattern 130. The data name group pattern 130 is, for example, a rule for extracting a plurality of data names related to the same type of devices installed in the same place, and extracting data names having different signal type categories as the same group. Is information defining a combination of categories assigned to the names of data to be analyzed. In other words, the data name set pattern 130 is information that defines a plurality of combinations in which the device type category and the installation location category are common to each other and the signal type categories are different from each other. Different signal type categories may be specified in advance by an analyst or the like, for example. For example, what is necessary is just to specify the signal classification category provided to the name of the data which should be the same analysis object. The data name set pattern 130 shown in FIG. 7 is a combination of categories for extracting the name of data for analyzing the supply air temperature at the time of valve opening in the air conditioner. Specifically, the data name set pattern 130 shown in FIG. 7 is a name of data related to an air conditioner installed in the same place, and a data name to which a category of “temperature, air supply, measurement” is given, “ Extract the data name given the category of “temperature, air supply, setting”, the data name given the category of “control output, valve opening”, and the data name assigned the category of “control output, state” Definition for The data name set pattern 130 includes, for example, definitions (α) to (δ), and the definitions (α) to (δ) are configured by combinations of variables A, B, C1, C2, D1, D2, and the like. Each variable is one of an installation location category, a device type category, and a signal type category. For example, the variable A indicates the installation location category, the variable B indicates the device type category, and the variables C1, C2, D1, E1, etc. indicate the signal type category. For example, the definition (α) includes a variable A [place owner], a variable B [air conditioner: machine owner], a variable C1 [temperature: believer], a variable D1 [supply air: Shinshu], and a variable E1 [measurement: trust. It is composed of a combination of [Osamu]. Similarly, the other definitions (β) to (δ) are configured by combinations of variables corresponding to categories. The variables A and B are common in the definitions (α) to (δ). The variable may be designated in advance by an analyst or the like.

  The extraction unit 18 extracts data names corresponding to the definitions (α) to (δ) of the data name group pattern 130 shown in FIG. That is, among the plurality of data names, a data name whose category matches the category of each definition is extracted. At this time, the names of data related to the same type of devices installed at the same place are extracted as the same set. For example, in the data name list with category 120 shown in FIG. 6, the data names corresponding to the definition (α) are the data name (1) and the data name (11). ) And the data name (11) are extracted. That is, since the category assigned to the data name (1) and the data name (11) matches the category constituting the definition (α), the extraction unit 18 extracts the data names (1) and (11). To do. Since the data name (1) and the data name (11) are data names related to the air conditioner, but the installation location is different, the extraction unit 18 sets the data name (1) and the data name (11) in different groups. Classify into: Since the data names corresponding to the definition (β) are the data name (6) and the data name (17), the extraction unit 18 extracts the data name (6) and the data name (17). Since the data name (6) and the data name (17) have different installation locations, the extraction unit 18 classifies the data name (6) and the data name (17) into different sets. Similarly, the definitions (γ) and (δ) are classified by extracting corresponding data names. The extraction unit 18 corresponds to the definitions (α) to (δ) regardless of the order of variables included in the definitions (α) to (δ) and the order of terms (categories) included in the data names. Extract the data name.

  FIG. 8 shows the extraction result. As shown in FIG. 8 (a), the data names (1), (6), (4), (5) are the names of data related to the air conditioners installed at the same place. These data names are classified into the same data name set 1. Further, as shown in FIG. 8B, the data names (11), (17), (15), and (16) are names of data related to the air conditioners installed in the same place, and therefore the extraction unit 18 Classifies these data names into the same data name set 2.

  Then, the extraction unit 18 outputs the data name sets 1 and 2 as the analysis target data name set 140, respectively. The data extraction device 230 illustrated in FIG. 1 extracts data to which the data name included in the data name set 1 is assigned from the storage device 222, and the extracted data is the analysis target data 240 related to the data name set 1. Output to. Similarly, the data extraction device 230 extracts data to which the data name included in the data name set 2 is assigned from the storage device 222, and extracts the extracted data as analysis target data 240 related to the data name set 2 to the analysis device 250. Output. Then, the analysis device 250 performs diagnosis and analysis related to the supply of the air conditioner installed in B1F of the building 200 based on the analysis target data related to the data name set 1. Similarly, the analysis device 250 performs diagnosis and analysis related to air supply of an air conditioner installed in 2F of the building 200 based on the analysis target data related to the data name set 2.

  As described above, a category is assigned to a data name using a category dictionary, and data names relating to devices having the same installation location are extracted using a data name set pattern 130 in which a set of data names to be extracted is defined by category. Thus, it is possible to classify data names that could not be classified only by the similarity of terms. For example, when the data names are classified only by the similarity of terms, “B1F system air conditioner SA temperature”, “B1F system air conditioner RA temperature” and “2F system air conditioner SA temperature” have only one term. Similarities are the same in the sense that they are different, and these data names cannot be distinguished and extracted. On the other hand, in this embodiment, by assigning a signal type category to a data name, it becomes possible to distinguish terms related to a signal type such as “SA” and “RA”, and to install in different places. By classifying the data names relating to the selected devices into different groups, it becomes possible to classify data names that cannot be classified only by the similarity of terms. Since signal modifiers such as “SA (supply air)” and “RA (return air)” can also be distinguished from each other, it is possible to distinguish and extract data names including the signal modifier. As a result, it is possible to extract only the names of data useful for data analysis. For example, data having a data name including “SA (supply air)” and data having a data name including “RA (return air)” are subject to analysis, so that the data name including “SA” and “RA” If the data names including are classified into the same set, the data analysis may be hindered and the accuracy of the data analysis may be reduced. In the present embodiment, data names having different signal modifiers can be classified into different groups, so that the accuracy of data analysis can be improved. As described above, since the names of data to be analyzed can be automatically classified, it is possible to save labor of the analyst and reduce the time required for extracting data to be analyzed.

(Modification 1)
Next, Modification 1 will be described. In the above-described embodiment, a case has been described in which a data name that completely matches the definition (for example, each of the definitions (α) to (δ)) included in the data name set pattern 130 is extracted from the category-attached data name list 120. . In the first modification, even if the definition included in the data name combination pattern 130 does not completely match, a data name having a high matching degree (matching degree) is extracted. For example, the extraction unit 18 uses the category-specific data name as the data name to which the same category as some of the device type category, signal type category, and installation location category defined by the data name set pattern 130 is assigned. Extract from list 120.

  FIG. 9 shows an example of the data name list 100A according to the first modification. The data name list 100A includes the data names (1) to (19) as in the data name list 100 shown in FIG. 3, but the data name list 100A includes the data names (1) to (3) and (11) to (13). The term “measurement” is not included in the data name. This point is different between the data name list 100 and the data name list 100A.

  Similar to the above-described embodiment, the morphological analysis unit 12 classifies each data name included in the data name list 100A by each term, and the category assigning unit 14 assigns a category to each term included in the divided data name. To do. Thereby, a data name list to which a category is assigned as shown in FIG. 6 is created. In the category-attached data name list according to the first modification, the data names (1) to (3) and (11) to (13) are related to “measurement”, unlike the category-attached data name list 120 shown in FIG. The signal type category (signal modifier category) is not assigned.

  In this case, the extraction unit 18 compares the data name group pattern 130 with the category-added data name list according to the first modification, and selects data names corresponding to the definitions (α) to (δ) of the data name group pattern 130. Then, the data is extracted from the category-added data name list according to the first modification.

  FIG. 10A shows an example of the data name set pattern 130. The pattern shown in FIG. 10A is the same as the pattern shown in FIG. In Modification 1, since the term “measurement” is not included in each data name, the data name that matches the definition (α) including the signal type category related to “measurement” among the definitions (α) to (δ) is Only data names that match each of the definitions (β) to (δ) are extracted without being extracted.

  FIG. 10B shows a set of data names (data name set 1) that matches each of the definitions (β) to (δ). Similarly to the data name set 1 shown in FIG. 8A, the data names (4) to (6) are extracted and classified into the same data name set 1.

  FIG. 10C shows a data name set (data name set 2) that matches each of the definitions (β) to (δ). Similarly to the data name set 2 shown in FIG. 8B, the data names (15) to (16) are extracted and classified into the same data name set 2.

  In the first modification, the extraction unit 18 extracts the data name closest to the definition (α) from the category-attached data name list. Focusing on candidate data names included in the data name set 1, the definition (α) includes the variables [A: [field]], variables [B: [air conditioner: machine owner]], and variables [C1: [temperature]. : [Shinsha]], variable [D1: [supply: Shinshu]], variable [E1: [measurement: Shinshu]], data name (1) to (10) of the data name ( Since 1) is “B1F system air conditioner SA temperature”, the matching degree (matching degree) of the data name (1) with respect to the definition (α) is the highest among the data names (1) to (10). Therefore, the extraction unit 18 extracts the data name (1) and classifies it into the data name set 1. Similarly, when attention is paid to candidate data names included in the data name set 2, the degree of match (degree of coincidence) of the data name (11) with respect to the definition (α) among the data names (11) to (19). Highest. Therefore, the extraction unit 18 extracts the data name (11) and classifies it into the data name set 2.

  As described above, even if there is no data name that completely matches the data name set pattern 130 that is the basis of extraction, the data name to be extracted can be extracted by considering the matching degree. It becomes possible.

(Modification 2)
Next, Modification 2 will be described. In the second modification, the extraction unit 18 specifies the data type based on the value of the data itself associated with the data name and the unit attached to the data, and the data is based on the signal type category corresponding to the specified type. Extract names. For example, when there is no data name that completely matches the definition included in the data name set pattern 130, the signal type is identified by referring to the value or unit of the data itself, and the signal type category corresponding to the identified signal type Is added to the data name to classify the data name. Alternatively, even if there is a data name that completely matches the definition included in the data name set pattern 130, the signal type is specified by referring to the value or unit of the data itself, and based on the specified signal type Data name extraction may be performed by correcting the signal type category. By correcting the signal type category in this way, for example, it is possible to prevent the data name from being erroneously extracted.

  FIG. 11A shows an example of a unit assumed to be given to data. For example, if the unit given to the data is “° C.”, the type of the data is estimated to be “temperature”. Therefore, the extraction unit 18 estimates the data to which the unit “° C.” is assigned as temperature data, and performs the extraction process by collating the name of the data with the data name set pattern 130. For example, the extraction unit 18 acquires data with a data name that does not match the definition included in the data name set pattern 130 from the storage device 222, and estimates the data with the unit “° C.” as temperature data and performs extraction processing. I do. As a result, even if the data name does not include the term “temperature”, if the unit of the data to which the data name is assigned is “° C.”, the data is estimated to be temperature data, and the data name The extraction process can be performed as including the term “temperature”. For example, when a term such as “Temp” is given to a data name as a kind of temperature term, and the signal type category related to temperature is not given to the data name, the data name is used as the data name of “temperature”. Although it cannot be extracted, according to the second modification, if “° C.” is included in the data unit, the data name can be extracted as the data name of “temperature”.

  Even if the data name includes a term of a physical quantity different from “temperature” such as “pressure”, if the data unit is “° C.”, the extracting unit 18 Is extracted as temperature data. Thereby, even if the data name is incorrect, the data name can be accurately extracted. For example, the extraction unit 18 corrects the signal type category related to “pressure” given to the data name to the signal type category related to “temperature”, and performs extraction processing based on the corrected signal type category.

  Similarly, if the unit given to the data is “Pa”, it is estimated that the type of the data is “pressure”, and if the unit is “V”, the type of the data is “voltage”. If the unit is “A”, the type of data is estimated to be “current”. Based on the estimation, the extraction unit 18 collates the data name with the data name combination pattern 130 and performs a data name extraction process.

  FIG. 11B shows data (values). For example, data is acquired at regular intervals. The extraction unit 18 acquires data from the storage device 222, estimates the type of data based on the characteristics of the data, and classifies the data names based on the estimation. For example, the data type may be estimated based on the data range. Since the temperature itself, the pressure value, the current value, and the voltage value have different values and different ranges, the data type is estimated based on the ranges. Alternatively, if the data changes with time, the type of data may be estimated based on changes over a certain period, such as seasonal fluctuations, daily fluctuations, and monthly fluctuations. Alternatively, if the data does not change for a certain period, it may be estimated that it is a set value instead of a measured value. This is because it is considered that the measured value changes with time, and the set value does not change with time. As described above, by classifying the data names based on the characteristics of the data, the data names can be classified even when there is no data name that completely matches the definition included in the data name set pattern 130. . Even if the data name is incorrect, the data name can be accurately extracted.

  As another example, the extraction unit 18 may estimate the location where the device is installed based on the characteristics of the data. Since the data change tendency may vary depending on the installation location of the device, the extraction unit 18 may estimate the installation location of the device based on the change tendency. For example, if the data change tendency in which the installation name term is not included in the data name is the same as the data change tendency in which the installation location of the device is already known, the extraction unit 18 identifies the installation location as the data name. It is presumed that this is a place where data that does not include the installation location term is acquired. For example, past data may be used as data where the installation location of the device is known. Then, the extraction unit 18 adds an installation location category corresponding to the estimated installation location to the data name, and extracts the data name. As an example, when the tendency of changes in measured values such as temperature, pressure, and humidity differs depending on the installation location of equipment in the building 100, the installation location of equipment in the building 100 is estimated based on the tendency of change in measured values. The data name extraction process is performed. Further, when the setting value varies depending on the installation location of the device in the building 100, the installation location of the device in the building 100 may be estimated based on the setting value, and the data name extraction process may be performed.

  The data name extraction device 10 described above includes a processor such as a CPU (not shown). A processor implement | achieves each function of the morphological analysis part 12, the category provision part 14, and the extraction part 18 by running the program memorize | stored in the memory which is not shown in figure. The program is stored in a storage device such as a hard disk drive (HDD) via a recording medium such as a CD or DVD, or via a communication path such as a network. The program may be stored in advance in a storage device such as a hard disk drive. A function stored in a storage device such as a hard disk drive is read into a memory such as a RAM and executed by a processor such as a CPU, thereby realizing the functions of each unit of the data name extraction device 10.

  10 data name extraction device, 12 morpheme analysis unit, 14 category assigning unit, 16 category storage unit, 18 extraction unit.

Claims (6)

It consists of device type terms related to the type of device to be detected by the sensor, signal type terms related to the type of sensor output value, and installation location terms related to the location where the device is installed, and each output value of multiple sensors Data name information associated with each data name information, assigning the same device type category to the device type terms that are synonymous with each other among the device type terms of each data name information, Category assigning means for assigning the same signal type category to the signal type terms that are synonymous with each other, and for assigning the same installation location category to the installation location terms that are synonymous with each other among the installation location terms of each data name information,
Obtains analysis target information that defines a combination of device type category, signal type category, and installation location category for making the output values of multiple sensors the same analysis target, and a plurality of data to which the category is given by the category assigning means Extraction means for extracting from the name information the data name information to which the same category as the category defined by the analysis target information is assigned;
A data name extraction device characterized by comprising: The data name extraction device according to claim 1,
The analysis target information defines a plurality of combinations in which the device type category and the installation location category are common to each other and the signal type categories are different from each other,
The extraction means extracts a plurality of data name information in which the device type category and the installation location category are common to each other and the signal type categories are different from each other, and the location indicated by the installation location term is the same among the extracted plurality of data name information. Classify certain data name information into the same set,
A data name extraction apparatus characterized by that. The data name extraction device according to claim 1 or 2,
When the data name information to which the same category as all categories of the device type category, the signal type category, and the installation location category defined by the analysis target information is not present in the plurality of data name information, Data name information to which the same category as some of the device type category, signal type category and installation location category defined by the analysis target information is extracted from a plurality of data name information.
A data name extraction apparatus characterized by that. The data name extraction device according to any one of claims 1 to 3,
The extraction means specifies the type of the output value based on the unit of the output value of the sensor, and extracts data name information based on the signal type category corresponding to the specified type.
A data name extraction apparatus characterized by that. The data name extraction device according to any one of claims 1 to 3,
The extraction means specifies the type of the output value based on the output value of the sensor, and extracts data name information based on the signal type category corresponding to the specified type.
A data name extraction apparatus characterized by that. On the computer,
It consists of device type terms related to the type of device to be detected by the sensor, signal type terms related to the type of sensor output value, and installation location terms related to the location where the device is installed, and each output value of multiple sensors Data name information associated with each data name information, assigning the same device type category to the device type terms that are synonymous with each other among the device type terms of each data name information, Assigning the same signal type category to the signal type terms that are synonymous with each other, and assigning the same installation location category to the installation location terms that are synonymous with each other among the installation location information of each data name information,
Obtain analysis target information that defines a combination of device type category, signal type category and installation location category for making the output values of multiple sensors the same analysis target, and from multiple data name information to which the category is assigned, Extracting data name information to which the same category as the category defined by the analysis target information is assigned;
A program characterized by having executed.

Images