WO2018088040A1 - マッチング装置、マッチング方法及びプログラム - Google Patents

マッチング装置、マッチング方法及びプログラム Download PDF

Info

Publication number
WO2018088040A1
WO2018088040A1 PCT/JP2017/034018 JP2017034018W WO2018088040A1 WO 2018088040 A1 WO2018088040 A1 WO 2018088040A1 JP 2017034018 W JP2017034018 W JP 2017034018W WO 2018088040 A1 WO2018088040 A1 WO 2018088040A1
Authority
WO
WIPO (PCT)
Prior art keywords
sensor
word
side metadata
application
matching
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2017/034018
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
紘 今井
哲二 大和
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Omron Corp
Original Assignee
Omron Corp
Omron Tateisi Electronics Co
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Omron Corp, Omron Tateisi Electronics Co filed Critical Omron Corp
Priority to US16/345,756 priority Critical patent/US10922325B2/en
Priority to EP17869161.4A priority patent/EP3540613B1/en
Publication of WO2018088040A1 publication Critical patent/WO2018088040A1/ja
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2457—Query processing with adaptation to user needs
    • G06F16/24578—Query processing with adaptation to user needs using ranking
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/907—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/23—Updating
    • G06F16/2379—Updates performed during online database operations; commit processing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/248—Presentation of query results

Definitions

  • the present invention provides a sensor that provides sensing data detected by a sensor and an application that provides services using the sensing data by using the respective metadata to perform sensing that satisfies the requirements of the application.
  • the present invention relates to a technique for extracting candidate sensors capable of providing data.
  • sensor network A technology called sensor network is being studied. This is because sensor devices with sensing and communication functions (hereinafter also simply referred to as “sensors”) are installed in various locations and industrial facilities, and networking them to collect, manage, and seamlessly sense data. It can be used.
  • sensors are installed to collect data that the owners themselves need. Therefore, it is often not used except when the owner collects data (the sensor itself is not operating, or sensing data is not used even if the sensor is operating). For this reason, the distribution of sensing data is low, and no matter how meaningful the data is for a third party, the sensor owner himself has only analyzed and used it. As a result, redundant investment of equipment and network congestion due to communication with sensors installed by each person have been invited.
  • IoT Internet of Things
  • sensing data collected in large quantities all over the world according to their own purposes For example, processing data on an application server capable of processing big data to create and provide added value, or activating trading of sensing data to bring about an economic effect.
  • the owner of the sensor can obtain a price by allowing the data user to temporarily use the sensor or providing sensing data. For the user, there is an advantage that necessary data can be obtained at low cost because investment for installing the sensor is unnecessary.
  • Patent Document 1 an invention relating to a mechanism for appropriately distributing resources such as sensing data has been proposed.
  • mapping is performed between application-side metadata and sensor-side metadata, and an application that requires sensing data is associated with a sensor that can provide the data. Then, a data flow control command is transmitted to the device that manages the sensor.
  • metadata refers to information used for searching and matching by the server
  • sensor-side metadata is information related to the attributes of the sensor and sensing data obtained by the sensor
  • application-side metadata is required for the application itself and the application. This refers to information related to the attributes of sensing data.
  • the data flow control command is command information that includes information for specifying a sensor that is a data providing source and an application that is a data using destination, and commands the data to be distributed from the data providing source to the data using destination. .
  • the words registered in the sensor-side metadata and the application-side metadata are not uniform and vary to some extent.
  • the temperature sensor may be described as a temperature sensor, or may be described as a thermometer or the like.
  • various expressions such as fluctuations in expression, synonyms, synonyms, and upper or lower concepts can be considered.
  • the present invention has been invented in view of the above-described problems, and provides a technique for increasing the chance of matching between sensor-side metadata and application-side metadata and suppressing a decrease in matching accuracy. is there.
  • the present invention relates to a sensor-side metadata acquisition unit that acquires sensor-side metadata that is information related to a sensor that outputs sensing data, and application-side metadata that is information related to an application that provides a service using the sensing data.
  • the acquired application-side metadata acquisition unit, the first related word related to the acquired sensor-side metadata word, and the second related word related to the acquired application-side metadata word are common.
  • a sensor management apparatus that manages the sensor, and a matching unit that performs matching between the sensor-side metadata and the application-side metadata, and extracts sensor candidates that can provide sensing data that satisfies the application requirements Extracted by the matching unit
  • a matching device including an instruction unit for transmitting a data flow control command which includes a sensor and information identifying the application.
  • the intention of data distribution of the user on the sensor side or the application side can be reflected, the chance of matching between the sensor side metadata and the application side metadata can be increased, and a decrease in matching accuracy can be suppressed. can do.
  • the chances of matching can be increased, and a decrease in matching accuracy can be suppressed.
  • an appropriate amount of data flow control commands can be output, and network communication traffic can be suppressed. Can effectively use the resources of the entire system.
  • the matching unit has at least a first relationship between the word on the sensor side metadata and the first related word or a second relationship between the word on the application side metadata and the second related word.
  • a relevance calculating unit for calculating one of the first related word and the first relevance, or at least one of the second related word and the second relevance;
  • a sensor candidate extraction unit that performs matching between metadata and the application-side metadata and extracts sensor candidates that can provide sensing data that satisfies the application requirements may be provided.
  • the degree of matching between sensor-side metadata and application-side metadata can be indexed by introducing the relationship between words and related words.
  • the sensor candidate extraction unit may use the first related word and the first related word in which the first relevance is within a predetermined range, or the second related word in which the second relevance is within a predetermined range.
  • the sensor-side metadata and the application-side metadata may be matched using the second relationship.
  • the increase in the related word common to the word of a sensor side metadata and the word of an application side metadata will be suppressed by using the related word whose relevance is in a predetermined range for matching. As well as more accurate matching.
  • the relevance range to be used can be set to a desired relevance range, and the number of related words used for matching can be adjusted.
  • the range setting unit uses the relevance range of the related word at the time of the contract establishment of the related word relevance used in the next matching process. You may make it the structure set as a range.
  • the range of related words can be set to an appropriate range by using the related range of related words at the time of contract establishment in the next matching process.
  • the sensor candidate extraction unit may correct the first relevance or the second relevance in consideration of co-occurrence of words of the sensor side metadata or the application side metadata.
  • the words which co-occur in the said sensor side metadata or the said application side metadata are words with high relevance, and correct
  • word co-occurrence is also one of the relevance indicators of the present invention.
  • the sensor candidate extraction unit may use the sensor-side metadata word and the first related word or the application-side metadata word and the second related word in the past for matching.
  • the first relevance or the second relevance may be corrected according to the frequency of establishment of the contract.
  • the word database is updated with the word of the sensor side metadata or the application side metadata used for matching, the first or second related word, or the newly calculated first or second relation.
  • the word database updating unit may be configured.
  • a sensor candidate presentation unit that presents information regarding the extracted candidate sensors to the user on the application side may be used.
  • the matching method acquires sensor side metadata that is information about a sensor that outputs sensing data, acquires application side metadata that is information about an application that provides a service using the sensing data, and The sensor-side metadata and the application-side metadata in which the first related words related to the acquired sensor-side metadata word and the second related words related to the acquired application-side metadata word are common.
  • Data matching is performed, sensor candidates capable of providing sensing data that satisfies the application requirements are extracted, and the extracted sensor and the application are specified for a sensor management apparatus that manages the sensor.
  • Data flow control commands including information It is a method to trust.
  • the intention of data distribution of the user on the sensor side or the application side can be reflected, the chance of matching between the sensor side metadata and the application side metadata can be increased, and a decrease in matching accuracy can be suppressed. can do.
  • the chances of matching can be increased, and a decrease in matching accuracy can be suppressed.
  • an appropriate amount of data flow control commands can be output, and network communication traffic can be suppressed. Can effectively use the resources of the entire system.
  • the program according to the present invention includes processing for obtaining sensor-side metadata that is information relating to a sensor that outputs sensing data, and application-side metadata that is information relating to an application that provides a service using the sensing data.
  • the sensor-side metadata and the application-side in which the processing to acquire, the related words related to the acquired sensor-side metadata word, and the related words related to the acquired application-side metadata word are common
  • a process of extracting candidates for sensors capable of providing sensing data satisfying the requirements of the application by performing matching with metadata and for the sensor management device that manages the sensors, the extracted sensor and the application Data containing information to identify Is a program for executing a process of transmitting the flow control command to the computer.
  • the intention of data distribution of the user on the sensor side or the application side can be reflected, the chance of matching between the sensor side metadata and the application side metadata can be increased, and a decrease in matching accuracy can be suppressed. can do.
  • the chances of matching can be increased, and a decrease in matching accuracy can be suppressed.
  • an appropriate amount of data flow control commands can be output, and network communication traffic can be suppressed. Can effectively use the resources of the entire system.
  • the present invention can increase the chances of matching between the sensor-side metadata and the application-side metadata, and can suppress a decrease in matching accuracy.
  • FIG. 3 is a block diagram of a matching unit 13.
  • FIG. It is a figure for demonstrating operation
  • FIG. It is a figure for demonstrating the process of the sensor candidate extraction part.
  • 3 is a block diagram of a matching unit 13.
  • FIG. It is a figure for demonstrating operation
  • It is a block diagram of the matching part 13 in the modification of 1st Embodiment.
  • FIG. 1 It is a figure for demonstrating the modification 1 of 1st Embodiment. It is a figure for demonstrating the modification 2 of 1st Embodiment. It is a block diagram of the other modification of 1st Embodiment. 3 is a block diagram of a matching unit 13.
  • FIG. It is the figure which showed an example of the word database 6 which the word database update part 135 registers. It is a figure for demonstrating operation
  • a sensor side that provides sensing data detected by the sensor, and a service using the sensing data Are matched using the respective metadata, and sensor candidates that can provide sensing data that satisfies the application requirements are extracted.
  • sensing data is distributed from the sensor side to the application side when the sensor side metadata matches the application side metadata.
  • FIG. 1 is a block diagram of the first embodiment.
  • 1 is a matching device
  • 2 is a thesaurus
  • 3 is a sensor data providing system
  • 4 is a data distribution control device (Distributor)
  • 5 is an application system.
  • Thesaurus 2 is a systematic thesaurus / dictionary that classifies words (words, etc.) and organizes them according to upper / lower relationships, partial / whole relationships, synonyms, synonyms, co-occurrence relationships, etc. It is a database.
  • the matching device 1 and the thesaurus 2 are connected via a network will be described.
  • the matching device 1 may include the thesaurus 2.
  • the sensor data providing system 3 is a system that provides sensing data acquired by a sensor.
  • the data distribution control device (Distributor) 4 has a function of receiving sensing data from the sensor data providing system 3 and distributing the sensing data to the application system 5 in accordance with a data flow control command to be described later.
  • Application system 5 is a system that provides services using sensing data provided by sensor data providing system 4.
  • the matching device 1 includes a sensor-side metadata acquisition unit 11, an application-side metadata acquisition unit 12, a matching unit 13, and a data flow control command instruction unit 14.
  • the sensor-side metadata acquisition unit 11 acquires sensor-side metadata that is information related to a sensor that outputs sensing data.
  • the senor is a device that detects some physical quantity or its change and records or outputs it as sensing data.
  • Typical examples of the sensor include a position sensor (GPS), an acceleration sensor, and a pressure sensor, but a camera, a microphone, an input system, and the like can also be said to be sensors.
  • the sensor-side metadata refers to information regarding the sensor and attributes of sensing data obtained by the sensor, which are used for matching described later.
  • Examples of the sensor-side metadata items include data category, data item, measurement area, measurement date and time, device, distribution interval, and data specification.
  • the data category is “indoor environment measurement”
  • the data item is “temperature, humidity, sound pressure, acceleration”
  • the measurement area is “Shiga Prefecture”
  • the measurement date is “immediate”
  • the device is “device name”.
  • “one time / 10 minutes” is described in the distribution interval
  • “http://WWW.XXX” is described in the data specification.
  • nouns and phrases such as the number of times described in each item of the sensor side metadata are simply described as words of the sensor side metadata. When a plurality of words such as place names are gathered, the word may be decomposed into a plurality of meaningful words by morphological analysis or the like.
  • the sensor-side metadata acquisition unit 11 can take various configurations. For example, there is a method of storing these metadata in a database accessible from the matching device 1. In this case, since the metadata necessary for the matching process is already in the database, event notification may be performed from the sensor side or the application side as a trigger to be acquired. Furthermore, the structure which does not have a sensor side metadata database can also be taken. In this case, metadata is transmitted from the management device or application of the sensor that detected the event occurrence.
  • the application-side metadata acquisition unit 12 acquires application-side metadata that is information related to an application that provides a service using sensing data.
  • application-side metadata refers to information related to the attributes of sensing data required by the application itself and the application.
  • Examples of application-side metadata items include data category, data item, measurement area, measurement date and time, and device. And “Environmental measurement” is described in the data category, “Temperature and humidity” in the data item, “Shiga prefecture...” in the measurement area, “Immediate” in the measurement date and time, “Device name” in the device, etc.
  • a noun or a phrase such as the number of times described in each item of application-side metadata is simply described as a word of application-side metadata. When a plurality of words such as place names are gathered, the word may be decomposed into a plurality of meaningful words by morphological analysis or the like.
  • the application side metadata acquisition unit 12 can take various configurations. For example, there is a method of storing these metadata in a database accessible from the matching device 1. In this case, since the metadata necessary for the matching process is already in the database, event notification may be performed from the sensor side or the application side as a trigger to be acquired. Furthermore, the structure which does not have an application side metadata database can also be taken. In this case, the application side user obtains the application side metadata that is information satisfying the application request at the timing of searching for the desired sensor.
  • the matching unit 13 includes a related word (hereinafter also referred to as a first related word) related to the acquired sensor-side metadata word, and a related word (related to the acquired application-side metadata word (
  • a related word related to the acquired application-side metadata word
  • the sensor-side metadata and the application-side metadata that are common to each other may be described as a second related term, and sensor candidates that can provide sensing data that satisfies the application requirements are extracted.
  • the related term refers to a word that is in a higher or lower concept of a word of sensor-side metadata or application-side metadata, a word that is synonymous, or a word that is in a synonymous relationship.
  • the matching unit 13 searches the thesaurus 2 for related words related to the word of each item of the sensor side metadata. Similarly, the thesaurus 2 is searched for related words related to the word of each item of the application side metadata. If there is a related term that is common to both the sensor-side metadata and the application-side metadata among the retrieved related terms, matching is performed between the sensor-side metadata and the application-side metadata corresponding to the common related term. To extract sensor candidates that can provide sensing data that satisfies the application requirements.
  • the matching unit 13 extracts the sensor side metadata and the application side metadata used for matching in order to extract, as candidates, sensors that may satisfy the above-described application user requirements. , It expands to sensor-side metadata and application-side metadata where the related terms related to the words of the sensor-side metadata and the related terms related to the words of the application-side metadata are common, increasing the chances of matching and matching The reduction in accuracy is suppressed.
  • the related words related to the words of the sensor side metadata or the related words related to the words of the application side metadata are obtained, if the related words that are not so related are used, the words of the sensor side metadata and the application side metadata are used.
  • the number of related words that are common to data words increases, which not only imposes a burden on processing, but also may reduce matching accuracy. Therefore, if it is possible to limit the related words to be used after understanding how much the words in the sensor-side metadata or the application-side metadata relate to the related words, more accurate matching can be performed. .
  • FIG. 2 is a block diagram of the matching unit 13.
  • the matching unit 13 includes a relevance calculation unit 131 and a sensor candidate extraction unit 132 as shown in FIG.
  • the relevance calculation unit 131 is registered in the word and thesaurus 2 of each item of the sensor side metadata. That calculate the relevance (second relevance) between the related word (first relevance) and the word of each item in the application-side metadata and the related word registered in the thesaurus 2 It is.
  • relevance indicates how similar or related words are.
  • the conventional method can be used regardless of the type.
  • the relevance described in the literature Automatic extraction of related terms using search engines: AutomaticAutoExtraction of Related Terms using Web Search Engines, Keigo Watanabe, Danushka Bollegala, Yutaka Matsuo, Mitsuru Ishizuka
  • As another method for calculating relevance there is a method of expressing a position of a word in a word space as a distance. According to this method, the closer the distance between words, the higher the relevance between words, and the farther the distance between words, the lower the relevance between words.
  • the relationship calculation unit 131 as an example of the relationship between words, the value of how much the words differ from the viewpoint of the upper or lower concept, synonym relationship, synonym relationship, co-occurrence relationship This will be described using the degree of difference F for calculating.
  • the value of the degree of difference F is smaller as the difference between the words is smaller, and is larger as the difference between the words is larger. Therefore, if the difference F is small, the relevance between words is high, and if the dissimilarity F is large, the relevance between words is low.
  • the calculation method of the difference degree F can use a conventional method, description about a specific calculation method is abbreviate
  • search for related words related to the word of each item in the sensor side metadata from the words registered in the thesaurus 2. For example, from the words registered in the thesaurus 2, words that are in the upper or lower concept with words of sensor-side metadata, words that are synonymous or synonymous, words that occur frequently, etc. Search for the first related term.
  • the dissimilarity F of the related word with respect to the word of the sensor side metadata is calculated.
  • F word X-related word
  • the dissimilarity F of each related word a, b, c, g with respect to the word X of the sensor-side metadata has the following value.
  • a word related to each item of the application-side metadata is searched from words registered in the thesaurus 2. For example, from words registered in the thesaurus 2, words that are in the upper or lower concept with words of application-side metadata, words that are synonymous or synonymous, words that occur frequently, etc. Search for a second related term.
  • the related word difference F with respect to the application side metadata is calculated.
  • Degree F is expressed as F (word Y-related word).
  • the difference F between the related words a, d, e, f, and g with respect to the word Y in the application side metadata is assumed to be the following value.
  • the sensor candidate extraction unit 132 will be described.
  • the sensor candidate extraction unit 132 performs matching between the sensor side metadata and the application side metadata based on the difference F calculated by the relevancy calculation unit 131.
  • the sensor-side metadata and application-side metadata used for matching are the sensor side where the related terms related to the words of the acquired sensor-side metadata and the related terms related to the words of the acquired application-side metadata are common. Metadata and application-side metadata. That is, the sensor-side metadata and the application-side metadata have related terms with a difference F value for both the sensor-side metadata and the application-side metadata. Then, matching is performed using the related words and the difference F between the sensor-side metadata and the application-side metadata.
  • the sensor candidate extraction by the sensor candidate extraction unit 132 will be described. In the following description, it is assumed that the degree of difference F is calculated by the above-described relevance calculation unit 131.
  • all related words for which the degree of difference F has been calculated may be used for matching.
  • the number of related words increases, which not only increases the load of matching processing, but also uses related words with low relevance.
  • matching that is not intended by the user may be performed. Therefore, related words having a difference F of less than a predetermined threshold are used as related words used for matching of each metadata. That is, related words with a difference F of less than a predetermined threshold are determined to have a small difference and high relevance, related words with a difference F of a predetermined threshold or more are determined to have a large difference and low relevance, Related words with high relevance are used for matching.
  • FIG. 3 shows an example in which the threshold value for determining the related word to be used is 1. More specifically, those having a sensor-side metadata item word X and dissimilarity F less than 1 (inside the dotted circle) are related word a, related word b, and related word g. Therefore, the first related word used as the related word of the word X of the sensor side metadata is the related word a, the related word b, and the related word g. On the other hand, since the related word c has a difference F of 1 or more, it is not used for matching.
  • the related word a, the related word d, and the related word e are those in which the word Y and the dissimilarity F of the application side metadata item are less than 1 (inside the dotted circle). Accordingly, the related words used as the second related word of the word Y of the application side metadata item are the related word a, the related word d, and the related word e. On the other hand, the related word f and the related word g are not used for matching because the dissimilarity F is 1 or more.
  • the threshold value of the difference F used as the related word is common to both the sensor-side metadata and the application-side metadata. However, different threshold values may be set.
  • the sensor candidate extraction unit 132 matches the sensor-side metadata and the application-side metadata in which the first related word of the word of the sensor-side metadata and the second related word of the word of the application-side metadata are common.
  • the sensor of the sensor side metadata is extracted as a sensor candidate.
  • the first related words of the word X of the sensor side metadata are the related word a, the related word b, and the related word g.
  • the second related words of the word Y in the application-side metadata item are a related word a, a related word d, and a related word e.
  • a common related word is the related word a. Accordingly, it is determined that the sensor-side metadata and the application-side metadata that share the related term a are matched, and the sensor-side metadata sensor having the word X is extracted as a sensor candidate.
  • the difference F between the word X of the sensor-side metadata item and the related word is arranged in descending order on the left side, and the word Y and the related word of the application-side metadata item are arranged on the right side. Are arranged in descending order of value. And the example which uses the related word of the difference F whose difference F is less than the threshold value 1 for a matching is shown.
  • the related words having the word X and the difference F of less than 1 in the sensor-side metadata item are the related word a, the related word b, and the related word g. Therefore, the first related word used as the related word of the word X of the sensor side metadata is the related word a, the related word b, and the related word g. On the other hand, since the related word c has a difference F of 1 or more, it is not used for matching.
  • related words having a word F and a difference F of less than 1 in the application-side metadata item are related word a, related word d, and related word e. Accordingly, the related words used as the second related word of the word Y of the application side metadata item are the related word a, the related word d, and the related word e. On the other hand, the related word f and the related word g are not used for matching because the dissimilarity F is 1 or more.
  • the common related word is related word a. Accordingly, it is determined that the sensor-side metadata and the application-side metadata that share the related term a are matched, and the sensor-side metadata sensor having the word X is extracted as a sensor candidate.
  • the matching unit 13 may be configured to present to the user on the application side what has been extracted as a sensor candidate.
  • FIG. 5 is a block diagram of the matching unit 13 having a function of presenting candidate sensors that can provide sensing data that satisfies the requirements of the matched application to the user on the application side.
  • the matching unit 13 includes a sensor candidate presentation unit 133.
  • the sensor candidate presentation unit 133 presents the sensor candidate obtained from the sensor candidate extraction unit 132 and the metadata of the sensor to the user on the application side.
  • the presentation method may display the sensor candidate and the metadata of the sensor on a display, print out a document describing the contents, or send the document to the user by e-mail describing the contents. May be.
  • absolute pressure sensor In the following description, “absolute pressure sensor”, “blood pressure monitor”, and “load sensor” exist as sensors, and the sensor side metadata of each sensor is referred to as sensor side metadata 1 of “absolute pressure sensor”. Assume that there is an “absolute pressure sensor”, the word “sphygmomanometer” is present as sensor-side metadata 2 of “blood pressure monitor”, and the word “load sensor” is present as sensor-side metadata 3 of “load sensor”. Further, the application side metadata includes the word “pressure sensor”.
  • the relevance calculation unit 131 searches the thesaurus 2 for related words related to the word “absolute pressure sensor” in the sensor side metadata 1.
  • the related word “atmospheric pressure sensor” and the related word “temperature sensor” are retrieved from the thesaurus 2 as related words of the word “absolute pressure sensor” in the sensor side metadata 1.
  • the degree of difference F between the word “absolute pressure sensor” in the sensor side metadata 1 and the related word is calculated.
  • the calculation results are as follows.
  • the relevancy calculation unit 131 searches the thesaurus 2 for related words related to the word “blood pressure monitor” in the sensor side metadata 2.
  • the related word “pressure sensor”, the related word “blood pressure monitor”, and the related word “blood pressure sensor” are searched from the thesaurus 2 as related words related to the word “blood pressure monitor” in the sensor side metadata 2.
  • the degree of difference F between the word “blood pressure meter” in the sensor side metadata 2 and the related word is calculated. The calculation results are as follows.
  • Dissimilarity F word “blood pressure meter” ⁇ related word “pressure sensor”
  • Difference F word “blood pressure monitor” ⁇ related word “blood pressure monitor”
  • Difference F word “blood pressure meter” ⁇ related word “blood pressure sensor”
  • the relevance calculation unit 131 searches the thesaurus 2 for related words related to the word “load sensor” in the sensor side metadata 3.
  • the related word “strain gauge” and the related word “weight sensor” are searched from the thesaurus 2 as related words related to the word “load sensor” in the sensor side metadata 3.
  • the degree of difference F between the word “load sensor” in the sensor side metadata 3 and the related word is calculated.
  • the calculation results are as follows.
  • the relevancy calculation unit 131 searches the thesaurus 2 for related words related to the word “pressure sensor” in the application side metadata.
  • related words related to the word “pressure sensor” in the application side metadata from the thesaurus 2, the related word “pressure gauge”, the related word “pressure monitor”, the related word “pressure monitor”, and the related word “barometric pressure sensor”. ”, The related word“ blood pressure monitor ”, and the related word“ strain gauge ”.
  • the difference F between the word “pressure sensor” of the application-side metadata and the related word is calculated. The calculation results are as follows.
  • the sensor candidate extraction unit 132 includes related words related to both the sensor-side metadata and the application-side metadata among the related words (the value of the difference F with respect to both the sensor-side metadata and the application-side metadata). Matching is performed using a related word) and its dissimilarity F. At that time, matching is performed by limiting to related words having a difference F of less than 1.
  • related words whose difference F from the word “absolute pressure sensor” in the sensor-side metadata 1 is less than 1 are the related word “barometric sensor” and the related word “temperature sensor”. Further, related words having a difference F of less than 1 from the word “blood pressure monitor” in the sensor-side metadata 2 are related words “pressure sensor”, related words “blood pressure monitor”, and related words “blood pressure sensor”. Further, related words having a degree of difference F from the word “load sensor” of the sensor side metadata 3 of less than 1 are related words “strain gauge” and related word “weight sensor”.
  • related words whose difference F from the word “pressure sensor” in the application side metadata is less than 1 are related words “pressure gauge”, related words “pressure monitor”, related words “pressure monitor”, and related words “barometric pressure”. Sensor ”and related term“ blood pressure monitor ”.
  • the related terms common to the sensor side metadata and the application side metadata are the related term “barometric pressure sensor” and the related term “blood pressure monitor”.
  • the sensor candidate extraction unit 132 extracts “sphygmomanometer” and “absolute pressure sensor” as sensor candidates that can provide sensing data that satisfies the requirements on the application side.
  • priorities of the extracted “blood pressure monitor” and “absolute pressure sensor” may be assigned.
  • priorities of the extracted “blood pressure monitor” and “absolute pressure sensor” may be assigned.
  • the total value of “blood pressure monitor” and “absolute pressure sensor” is as follows.
  • the sensor candidate presentation unit 133 presents the sensor candidates to the user on the application side.
  • An example presented to the user on the application side is shown in FIG. In FIG. 7, “Sensor A” of “Sphygmomanometer” of the sensor side metadata extracted by the sensor candidate extraction unit 132, “Sensor B” of “Absolute pressure sensor” of the sensor side metadata, and their ranks and A list of differences F is presented.
  • the data flow control command instruction unit 14 receives the result of matching by the matching unit 13 and transmits a data flow control command to the data distribution control device 4.
  • This data flow control command includes information for specifying the sensor data providing system 3 that is the data providing source and the application system 5 that is the data using destination, and commands that data is distributed from the data providing source to the data using destination. Command information.
  • the data flow control command provides sensor data that provides the sensing data of the “blood pressure monitor” sensor when the application system 5 side designates a “blood pressure monitor” sensor with priority 1 in the example of FIG. 6 described above. It includes information that identifies the system 3 and the application system 5 that uses the sensing data of the sensor of the “sphygmomanometer” that registers the “pressure sensor” as metadata. This is command information for instructing to circulate the sensing data of the “blood pressure monitor” sensor.
  • the information specifying the sensor data providing system 3 and the application system 5 that is the data use destination includes, for example, an IP address.
  • the data flow control command can instruct the sensing data of a certain sensor to be distributed to a plurality of applications.
  • each sensor can distribute sensing data to one application.
  • indication part 14 is a trigger which outputs a data flow control command, there are the following cases.
  • the matching unit 13 determines that the sensor-side metadata matches the application-side metadata under a predetermined condition.
  • the sensor-side metadata is detected by the sensor candidate presentation unit 133 of the matching unit 13. If the application-side user wants to provide sensing data based on the matching result, the above-mentioned case is just an example. It is not limited only to the case.
  • the matching target is expanded to the sensor-side metadata and the application-side metadata having a common related word, so that the chances of matching can be increased. Therefore, it is possible to suppress a decrease in matching accuracy. Therefore, an appropriate amount of matching can be obtained while ensuring accuracy, and as a result, an appropriate amount of data flow control commands can be output, network communication traffic can be suppressed, and the entire system can be controlled. You can also effectively use these resources.
  • the intention of the user on the sensor side or the application side is taken into consideration by using a related word having high relevance (for example, in the above example, a related word having a dissimilarity F within a predetermined threshold).
  • a related word having high relevance for example, in the above example, a related word having a dissimilarity F within a predetermined threshold.
  • the possibility of matching is increased.
  • the number of related words to be referenced is reduced, it is possible to reduce the processing load required for matching.
  • matching is performed between the application-side metadata and the sensor-side metadata, and matching between an application that requires sensing data and a sensor that can provide the data is performed. Done. Then, a data flow control command is transmitted to the device that manages the sensor. As a result, the distribution of sensing data taking various conditions into consideration is promoted, the service is improved, and it is beneficial to both the data provider and the user.
  • both the relationship between the sensor-side metadata word and the related word and the relationship between the application-side metadata word and the related word are calculated. Only one of them may be calculated.
  • the word “blood pressure meter” of the sensor-side metadata 2 and the word “pressure sensor” of the application-side metadata can be said to be related terms that are common to each other. May be calculated (for example, the degree of difference F).
  • the value of the dissimilarity F decreases as the relevance between the words increases.
  • the relationship between the words and the words may be reduced.
  • the value of the difference F may increase.
  • the threshold value for selecting a related word used for matching is naturally different, and the value of the degree of difference F for the selected related word is also less than the threshold value.
  • the degree of difference F is an example for knowing the relationship between words, and it is possible to use other methods without being limited to the degree of difference F as long as the relationship between words can be known. That's right.
  • a threshold value of the difference F for selecting a related word to be used is set in the sensor candidate extraction unit 132 in advance.
  • related words used for matching cannot be extracted with the set threshold value. That is, the range of relevance of related words used for matching is narrow.
  • the matching unit 13 is provided with a relevance threshold value setting unit 134 that changes the threshold value of the dissimilarity F for selecting a related word used for matching. Then, in order to widen the relevance range of related words used for matching, the relevance threshold value setting unit 134 increases the threshold value of the dissimilarity F set in the sensor candidate extraction unit 132 (for example, the threshold value is set from 1). Change to 1.2).
  • FIG. 9 is a diagram showing a change in related words used for matching caused by a threshold change by the relevance threshold setting unit 134.
  • the difference F between the word X of the sensor-side metadata item and the related word is arranged in descending order on the left side, and the difference between the word Y of the application-side metadata item and the related word on the right side. F are arranged in descending order. And the example which uses the related word of the difference F whose difference F is less than the threshold value 1 for a matching is shown.
  • the degree F is the following value.
  • the degree F has the following value.
  • the related words whose difference F is less than 1 from the word X of the sensor side metadata are the related word a, the related word b, and the related word g. Therefore, the first related word used as the related word of the word X of the sensor side metadata is the related word a, the related word b, and the related word g.
  • the related word c since the related word c has a difference F of 1 or more, it is not used for matching.
  • related words having a difference F of less than 1 from the word Y in the application side metadata are a related word a, a related word d, and a related word e. Accordingly, the related words used as the second related word of the word Y of the application side metadata item are the related word a, the related word d, and the related word e. On the other hand, the related word f and the related word g are not used for matching because the dissimilarity F is 1 or more.
  • the difference F is less than 1, and there is no related term common to the word X of the sensor-side metadata item and the word Y of the application-side metadata item.
  • the threshold value of the difference F is changed from 1 to 1.2.
  • the case where the threshold value of the difference F is changed from 1 to 1.2 is the lower side of FIG.
  • the related word a of the word X of the sensor metadata item whose dissimilarity F is 1.10 is newly included as a related word used for matching It is.
  • the related word a is a related word common to the sensor side metadata and the application side metadata. Accordingly, it is determined that the sensor-side metadata and the application-side metadata that share the related term a are matched, and the sensor-side metadata sensor having the word X is extracted as a sensor candidate.
  • the application-side user can operate the relevance threshold setting unit 134, and the application-side user can increase the degree of difference in order to widen the relevance range.
  • a configuration in which the threshold value of F is set to be large is also possible.
  • the difference F is used as an example of the relationship, but the present invention is not limited to this.
  • the relevance threshold value setting unit 134 determines the relevance range set in the sensor candidate extraction unit 132, and the difference F threshold value. (For example, the threshold value is changed from 1 to 0.5), and the related word is set to an appropriate number.
  • FIG. 10 is a diagram showing changes in related words used for matching caused by changing the threshold value by the relevance threshold value setting unit 134.
  • the difference F between the word X of the sensor-side metadata item and the related word, the word M of the application A-side metadata item, the word N of the application B-side metadata item, and the application C The side metadata items are arranged in the order of words O in descending order of the difference F. And the example which uses the related word of the difference F whose difference F is less than the threshold value 1 for a matching is shown.
  • the difference F between the related words b, d, q with respect to the word N of the application A side metadata is as follows: Suppose that it became a value.
  • the related words whose difference F is less than 1 from the word X of the sensor side metadata are the related word b, the related word c, the related word d, the related word f, and the related word r. Accordingly, the first related words used as the related words of the word X of the sensor side metadata are the related word b, the related word c, the related word d, the related word f, and the related word r. On the other hand, the related word a has a dissimilarity F of 1 or more and is not used for matching.
  • related words having a difference F of less than 1 from the word M in the application A side metadata are a related word b, a related word d, and a related word r. Therefore, the related words used as the second related word of the word Y of the application A side metadata are the related word b, the related word d, and the related word r. On the other hand, the related word f is not used for matching because the dissimilarity F is 1 or more.
  • related words having a difference F of less than 1 from the word N in the application B side metadata are related words b and d. Therefore, the related words used as the second related word of the word N of the application B side metadata are the related word b and the related word d. On the other hand, the related word q is not used for matching because the dissimilarity F is 1 or more.
  • related words having a difference F of less than 1 from the word O of the application C side metadata are a related word c and a related word f. Accordingly, the related words used as the second related word of the word O of the application C side metadata are the related word c and the related word f. On the other hand, the related word w is not used for matching because the dissimilarity F is 1 or more.
  • the threshold value of the difference F is changed from 1 to 0.5.
  • the case where the threshold value of the difference F is changed from 1 to 0.5 is the lower side of FIG.
  • the related words used for matching are reduced to related words b, related words d, and related words f.
  • the related word d is the only related word common to the sensor-side metadata and the application-side metadata.
  • the sensor-side metadata and the application A-side metadata that share the related term d are matched, and the sensor of the sensor-side metadata having the word M is extracted as a sensor candidate.
  • the threshold value by changing the threshold value, the range of related words can be narrowed, the number of related words used for matching can be reduced, and matching accuracy can be improved.
  • the application-side user can operate the relevance threshold setting unit 134, and the application-side user can reduce the degree of difference in order to narrow the relevance range.
  • a configuration in which the threshold value of F is set to be small is also possible.
  • the degree of difference F is used as an example of the relevance, but the present invention is not limited to this.
  • the purpose of the relevance range of related words by the relevance threshold setting unit 134 is that metadata intended by the user is used for matching. Therefore, as a method of setting a threshold for determining the relevance range of related terms to be used, it is set when a contract for providing sensor data to an application is established by matching sensor-side metadata and application-side metadata. There is a method in which the relevance threshold value setting unit 134 stores the threshold value and sets it as an initial value of the threshold value of the difference F for selecting a related word to be used at the start of the next matching process. This is because, when matching between the sensor-side metadata and the application-side metadata is established, it is considered that the range relationship of the relationship at that time is appropriate. That is, it is considered that the threshold value of the degree of difference F that determines the relevance range is appropriate.
  • examples of when a contract for providing sensor data to an application is established include the following cases.
  • each device transmits a contract establishment signal to the relevance threshold setting unit 134.
  • the relevance threshold setting unit 134 The threshold value of the related word dissimilarity F at the time of matching when the contract is established is set as an initial value of the threshold value of the related word dissimilarity F used at the start of the next matching process.
  • FIG. 11 is a block diagram of another modification of the first embodiment.
  • the relationship between the word calculated by the relevance calculation unit 131 and the related word can be stored in a database, there is an advantage that the number of words for calculating the relevance decreases at the time of matching performed later.
  • a word database 6 is provided in which the relationship between the word calculated by the word relationship calculation unit 131 and the related word is converted into a database.
  • the word database 6 is provided separately from the matching device 1 and an example in which the matching device 1 and the word database 6 are connected via a network will be described.
  • a word database 6 may be provided.
  • the matching unit 13 includes a word database update unit 135 as shown in FIG.
  • the word database update unit 135 registers (updates) the word searched by the relevance calculation unit 131 and the related word and the degree of difference F as an example of the relevance calculated by the relevance calculation unit 131 in the word database 6. It has a function.
  • FIG. 13 is a diagram showing an example of the word database 6.
  • FIG. 13 shows an example in which a related word for each word and its difference F are stored in association with each other.
  • Such a configuration makes it possible to reduce matching processing performed later by creating a database of related words that have been searched once and calculated relevance.
  • words that are input in parallel among words that are input in metadata are input when the user recognizes that the words are related to each other.
  • application-side metadata such as “atmospheric pressure sensor, weather”.
  • users on the side of providing sensing data often input related words in metadata.
  • the relevance calculation unit 131 calculates relevance, focusing on words that co-occur in each item of sensor-side metadata and application-side metadata. Note that, in the second embodiment described below, an example of the relationship calculated by the relationship calculation unit 131 will be described using the degree of difference F as in the first embodiment.
  • the relevancy calculation unit 131 first calculates the difference F between a word and a related word, as in the first embodiment. Specifically, it is as follows.
  • the difference F between the word “pressure sensor” in application-side metadata and related words Difference F (word “pressure sensor” -related
  • the word “blood pressure monitor” and the word “barometric pressure sensor” are words that co-occur with the word “weather”, and the co-occurrence degree C between the word “blood pressure monitor” and the word “weather” is The co-occurrence degree C of the word “barometric pressure sensor” and the word “weather” is calculated.
  • the calculation results are as follows.
  • the sensor candidate extraction unit 132 includes related words related to both the sensor-side metadata and the application-side metadata among the related words (the degree of difference with respect to both the sensor-side metadata and the application-side metadata). Matching is performed using F or a related word to which the degree of difference F ′ is attached) and the degree of difference F or the degree of difference F ′.
  • the threshold value for determining the related word to be used when the threshold value for determining the related word to be used is 1, it has a dissimilarity F or dissimilarity F ′ of less than 1, and sensor-side metadata and application-side metadata.
  • the related word “barometric sensor” is the only related word for both data.
  • the sensor candidate extraction unit 132 is a sensor of the sensor-side metadata 1 having a relation between the related word “barometric pressure sensor” and the related word as a sensor candidate capable of providing sensing data that satisfies the application-side request. "Pressure sensor" is extracted.
  • the total value of the difference F or difference F ′ of the related words common to the sensor side metadata and the difference F or difference F ′ of the related words common to the application side metadata is as follows.
  • the sensor candidate presentation unit 133 presents the sensor candidates to the user on the application side. An example presented to the user on the application side is shown in FIG. FIG. 15 presents a list of “sensor B” of “absolute pressure sensor” of the sensor-side metadata extracted by the sensor candidate extraction unit 132, their respective ranks, and their differences.
  • the sensor presented to the user is only “sensor B” of “absolute pressure sensor”, and the difference value is also small. (Relevance is high), it can be seen that highly accurate matching is performed.
  • the word database update unit 135 registers the calculated relevance (for example, the difference F or the difference F ′) in the word database 6. Also good. In this case, the co-occurrence word co-occurring with the related word and the related word difference F ′ corrected (adjusted) in relation to the co-occurrence word are registered.
  • FIG. 16 is a diagram showing an example of the word database 6 registered or updated by the word database update unit 135.
  • a co-occurrence word that co-occurs with the related word is stored in association with each other.
  • the related word for each word also includes the related word difference F ′ corrected (adjusted) in relation to the co-occurrence word.
  • the example of correcting (adjusting) the relationship between the sensor-side or application-side metadata and the related word described above is not limited to an example that considers the co-occurrence frequency of the word.
  • the word database update unit 135 obtains the number of times the contract is established in association with the words and related words of the sensor side or application side metadata when the contract is established in the matching process from the sensor candidate extraction unit 132, and the word database 6 is registered.
  • FIG. 17 is a diagram showing an example of the word database 6 registered or updated by the word database update unit 135 in such a case.
  • the relevance calculation unit 131 refers to the number of contracts established in the word database 6 and corrects the relationship between the words to be higher as the number of contracts is increased (for example, the degree of difference F). Decrease the value). For example, in the example of FIG. 17, since the contract is established once using the terms “pressure sensor” and “barometric sensor”, compared with the difference F in the case where the number of contract establishment as shown in FIG. 16 is not considered. The degree of difference F is corrected (adjusted) from 0.20 to 0.10 so as to increase the relevance.
  • the frequency of contract establishment is also one of the relevance indicators of the present invention.
  • the related word a, the related word b, the related word c, and the related word g are retrieved from the thesaurus 2 as related words related to the word X of the sensor side metadata item.
  • the relevance calculation unit 131 calculates the difference F between the word X and the related word related to the word X.
  • the difference F between the word X and each related word is as follows.
  • the relevance calculation unit 131 searches the thesaurus 2 for related words related to the word Y of each item in the application-side metadata.
  • a related word a, a related word d, a related word e, a related word f, and a related word g are retrieved from the thesaurus 2 as related words related to the word Y of each item of application-side metadata.
  • the relevancy calculation unit 131 calculates the difference F between the word Y and the related word related to the word Y.
  • the difference F between the word Y and each related word is as follows.
  • the sensor candidate extraction unit 132 will be described.
  • the sensor candidate extraction unit 132 performs matching between the sensor-side metadata and the application-side metadata based on the word searched by the relevance calculation unit 131 and the calculated relevance.
  • the sensor-side metadata and application-side metadata used for matching are the sensor side where the related terms related to the words of the acquired sensor-side metadata and the related terms related to the words of the acquired application-side metadata are common. Metadata and application-side metadata.
  • the sensor-side metadata and the application-side metadata have related terms with a difference F attached to both the sensor-side metadata and the application-side metadata. Then, matching is performed using the related words and the difference F between the sensor-side metadata and the application-side metadata.
  • the sensor candidate extraction by the sensor candidate extraction unit 132 will be specifically described with reference to FIG. In the following description, it is assumed that the difference degree F is calculated by the word relevance calculation unit 131 described above.
  • the related word a and the related word g are related words in which the difference F is given to both the sensor side metadata and the application side metadata, that is, the sensor side metadata and the application side.
  • the sensor candidate extraction unit 132 matches the sensor-side metadata and the application-side metadata using the related word a, the related word g, and the difference F thereof. I do.
  • the total difference F between the sensor-side metadata word X and the application-side metadata word Y via the related term a is as follows.
  • the distance between the sensor-side metadata word X and the application-side metadata word Y via the related term g is as follows.
  • the matching determination method when there are a plurality of related terms common to the sensor-side metadata and the application-side metadata, the sensor-side metadata and the application-side metadata via each related term This is a method of determining that the sensor-side metadata and the application-side metadata are matched when the total value of the differences F is equal to or less than a predetermined threshold value.
  • the total value of the dissimilarities F between the sensor-side metadata word X and the application-side metadata word Y via the related word a is 1.40.
  • the matching determination method when there are a plurality of common related terms between the sensor-side metadata and the application-side metadata, the sensor-side metadata and the application-side metadata via each related term When the value of the difference F is the smallest (highly related) and the selected difference F is equal to or less than a predetermined threshold, the sensor-side metadata and the application-side metadata are matched. This is how to judge. Then, the sensor corresponding to the sensor-side metadata is extracted as a sensor candidate that can provide sensing data that satisfies the application requirements.
  • the value F of the difference F between the word X of the sensor side metadata via the related term a and the word Y of the application side metadata is 1.40
  • the sensor side via the related term g Since the value F of the difference F between the metadata word X and the application-side metadata word Y is 2.00, the sensor-side metadata and the application-side metadata via the related word a are selected.
  • the predetermined threshold is set to 1.50
  • the degree of difference F is 1.50 or less, it is determined that the sensor-side metadata and the application-side metadata match.
  • the sensor-side metadata and the application via the related word a since the value F of the difference F between the sensor-side metadata and the application-side metadata via the related word a is 1.40, the sensor-side metadata and the application via the related word a Judge that side metadata matches. Then, the sensor corresponding to the sensor-side metadata is extracted as a sensor candidate that can provide sensing data that satisfies the application requirements.
  • the method for determining matching is not limited to the method described above.
  • a threshold value is not provided, and the relationship between sensor-side metadata and application-side metadata via a related word can be calculated, that is, sensor-side metadata and application-side metadata having a common related word May be extracted as matching candidates, and the sensor-side metadata sensor may be extracted as a sensor candidate capable of providing sensing data that satisfies the application requirements.
  • absolute pressure sensor In the following description, “absolute pressure sensor”, “blood pressure monitor”, and “load sensor” exist as sensors, and the sensor side metadata of each sensor is referred to as sensor side metadata 1 of “absolute pressure sensor”. Assume that there is an “absolute pressure sensor”, the word “sphygmomanometer” is present as sensor-side metadata 2 of “blood pressure monitor”, and the word “load sensor” is present as sensor-side metadata 3 of “load sensor”. Further, the application side metadata includes the word “pressure sensor”.
  • the relevance calculation unit 131 searches the thesaurus 2 for related words related to the word “absolute pressure sensor” in the sensor side metadata 1.
  • the related word “atmospheric pressure sensor” and the related word “temperature sensor” are searched from the thesaurus 2 as related words related to the word “absolute pressure sensor” of the sensor side metadata 1.
  • the degree of difference F between the word “absolute pressure sensor” in the sensor side metadata 1 and the related word is calculated. The calculation results are as follows.
  • the relevancy calculation unit 131 searches the thesaurus 2 for related words related to the word “blood pressure monitor” in the sensor side metadata 2.
  • the related word “pressure sensor”, the related word “blood pressure monitor”, and the related word “blood pressure sensor” are searched from the thesaurus 2 as related words related to the word “blood pressure monitor” in the sensor side metadata 2.
  • the degree of difference F between the word “blood pressure meter” in the sensor side metadata 2 and the related word is calculated. The calculation results are as follows.
  • Dissimilarity F word “blood pressure meter” ⁇ related word “pressure sensor”
  • Difference F word “blood pressure monitor” ⁇ related word “blood pressure monitor”
  • Difference F word “blood pressure meter” ⁇ related word “blood pressure sensor”
  • the relevance calculation unit 131 searches the thesaurus 2 for related words related to the word “load sensor” in the sensor side metadata 3.
  • the related word “strain gauge” and the related word “weight sensor” are searched from the thesaurus 2 as related words related to the word “load sensor” in the sensor side metadata 3.
  • the degree of difference F between the word “load sensor” in the sensor side metadata 3 and the related word is calculated.
  • the calculation results are as follows.
  • the relevancy calculation unit 131 searches the thesaurus 2 for related words related to the word “pressure sensor” in the application side metadata.
  • related words related to the word “pressure sensor” in the application side metadata from the thesaurus 2, the related word “pressure gauge”, the related word “pressure monitor”, the related word “pressure monitor”, and the related word “barometric pressure sensor”. ”, The related word“ blood pressure monitor ”, and the related word“ strain gauge ”.
  • the difference F between the word “pressure sensor” of the application-side metadata and the related word is calculated. The calculation results are as follows.
  • the sensor candidate extraction unit 132 adds related words related to both the sensor-side metadata and the application-side metadata (difference F is added to both the sensor-side metadata and the application-side metadata. Matching is performed using the related word) and the difference F thereof.
  • the application-side metadata 1 and the sensor-side metadata 1 are related via a related term “barometric sensor” (related term common to “barometric pressure sensor”), and the application-side metadata 1 and sensor
  • the side metadata 2 is related through the related terms “blood pressure meter” and “pressure sensor” (related terms common to “blood pressure meter” and “pressure sensor”), and the application side metadata 3 and the sensor side metadata.
  • Data 2 is related through a related word “strain gauge” (related word that “strain gauge” is common).
  • the total difference F between the sensor-side metadata and the application-side metadata via the related terms is as follows.
  • the sensor candidate presentation unit 133 may present the contents to the user on the application side.
  • An example of presenting sensor candidates to the user on the application side is shown in FIG. In FIG. 20, “sensor A” of “blood pressure meter” of the sensor side metadata extracted by the sensor candidate extraction unit 132, “sensor B” of “absolute pressure sensor” of the sensor side metadata, and the sensor side metadata
  • a list of “sensor C” of “load sensor” and a total value of respective ranks and degrees of difference F is presented.
  • the user on the application side can browse a list of presented sensor candidates and select a desired sensor.
  • Sensor side metadata acquisition unit 11 acquires sensor side metadata (Step 100).
  • the relevance calculation unit 131 searches the thesaurus 2 or the word database 6 for related words related to the words of the respective items in the sensor side metadata (Step 101).
  • the relevancy calculation unit 131 calculates the difference F (including correction of the difference F) between the sensor-side metadata and the related word (Step 102).
  • the relevance calculation unit 131 does not need to calculate the difference F when the difference F can be obtained from the word database 6.
  • Application side metadata acquisition unit 12 acquires application side metadata (Step 103).
  • the relevancy calculation unit 131 searches the thesaurus 2 or the word database 6 for related words related to the words of the respective items in the application-side metadata (Step 104).
  • the relevancy calculation unit 131 calculates the degree of difference F (including correction of the degree of difference F) between the application-side metadata and the related word (Step 105). The relevance calculation unit 131 does not need to calculate the difference F when the difference F can be obtained from the word database 6.
  • the sensor candidate extraction unit 132 performs matching processing between the sensor-side metadata and the application-side metadata based on the related word searched by the relevance calculation unit 131 and the calculated difference F (Step 106).
  • the specific matching process is as described above.
  • the data flow control command instructing unit 14 When the sensor side metadata and the application side metadata are matched (Step 107 Yes), the data flow control command instructing unit 14 outputs a data flow control command (Step 108).
  • the data flow control command instructing unit 14 is a trigger for outputting a data flow control command. As described above, there are the following cases.
  • the matching unit 13 determines that the sensor-side metadata matches the application-side metadata under a predetermined condition.
  • the sensor-side metadata is detected by the sensor candidate presentation unit 133 of the matching unit 13. If the application-side user wants to provide sensing data based on the matching result, the above-mentioned case is just an example. It is not limited only to the case.
  • Step 109 the process of matching failure is performed (Step 109).
  • the process for failure of matching is as follows.
  • Step 200 It is determined whether or not there is another process to be applied to the process for which matching is not established (Step 200). If any of the following processing is applicable, that processing is performed.
  • Modification 1 and Modification 2 of the first embodiment The threshold value (range) of the dissimilarity F used when selecting a related word used for matching is changed (Step 201).
  • Second Embodiment The difference F of related words is corrected (adjusted) in consideration of the co-occurrence of words in the sensor-side metadata or application-side metadata, the number of contracts, etc. (Step 202).
  • the threshold value (range) of the total value of the dissimilarity F determined to match the sensor side metadata and the application side metadata is changed (Step 203).
  • At least one of the above processes is performed, and the matching process between the sensor side metadata and the application side metadata is performed again (Step 106).
  • Step 202 the example of performing the second embodiment (Step 202) has been described as another process to be applied to the matching failure process.
  • the first matching process Step 106
  • related words are used.
  • the degree of difference F may be corrected (adjusted).
  • the relevance is not limited to the difference F, and any relevance may be used as long as it shows the relevance between words.
  • the matching device 1 includes, for example, a processor 201, a memory (ROM or RAM) 202, a storage device (hard disk, semiconductor disk, etc.) 203, an input device (keyboard, mouse, touch panel, etc.) 204, A general-purpose computer having hardware resources such as the display device 205 and the communication device 206 can be used.
  • the function of the matching device 1 is realized by a program stored in the storage device 203 being loaded into the memory 202 and executed by the processor 201.
  • the matching device 1 may be configured by a single computer or by distributed computing using a plurality of computers. Further, in order to increase the processing speed, part or all of the functions of the matching device 1 can be realized using dedicated hardware (for example, GPU, FPGA, ASIC, etc.).
  • the processor is Get sensor-side metadata, which is information about the sensor that outputs sensing data, Obtaining application-side metadata that is information about an application that provides services using the sensing data;
  • the sensor-side metadata and the application-side metadata in which the first related words related to the acquired sensor-side metadata word and the second related words related to the acquired application-side metadata word are common.
  • Perform matching with data extract sensor candidates that can provide sensing data that meets the requirements of the application,
  • a matching device that transmits a data flow control command including information that identifies the extracted sensor and the application to a sensor management device that manages the sensor.
  • (Appendix 2) Computer Get sensor-side metadata, which is information about the sensor that outputs sensing data, Obtaining application-side metadata that is information about an application that provides services using the sensing data; The sensor-side metadata and the application-side metadata in which the first related words related to the acquired sensor-side metadata word and the second related words related to the acquired application-side metadata word are common. Perform matching with data, extract sensor candidates that can provide sensing data that meets the requirements of the application, A matching method for transmitting a data flow control command including information that identifies the extracted sensor and the application to a sensor management device that manages the sensor.
  • (Appendix 3) A process of obtaining sensor-side metadata that is information related to a sensor that outputs sensing data; Processing for obtaining application-side metadata that is information related to an application that provides a service using the sensing data; Matching between the sensor-side metadata and the application-side metadata in which the related terms related to the acquired sensor-side metadata word and the related terms related to the acquired application-side metadata word are common Performing a process of extracting sensor candidates capable of providing sensing data satisfying the application requirements; A recording medium storing a program that causes a computer to execute a process of transmitting a data flow control command including information that identifies the extracted sensor and the application to a sensor management device that manages the sensor.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Library & Information Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
PCT/JP2017/034018 2016-11-14 2017-09-21 マッチング装置、マッチング方法及びプログラム Ceased WO2018088040A1 (ja)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US16/345,756 US10922325B2 (en) 2016-11-14 2017-09-21 Matching device, matching method, and program
EP17869161.4A EP3540613B1 (en) 2016-11-14 2017-09-21 Matching device, matching method, and program

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2016-221733 2016-11-14
JP2016221733A JP6406335B2 (ja) 2016-11-14 2016-11-14 マッチング装置、マッチング方法及びプログラム

Publications (1)

Publication Number Publication Date
WO2018088040A1 true WO2018088040A1 (ja) 2018-05-17

Family

ID=62110592

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2017/034018 Ceased WO2018088040A1 (ja) 2016-11-14 2017-09-21 マッチング装置、マッチング方法及びプログラム

Country Status (4)

Country Link
US (1) US10922325B2 (enExample)
EP (1) EP3540613B1 (enExample)
JP (1) JP6406335B2 (enExample)
WO (1) WO2018088040A1 (enExample)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180157692A1 (en) * 2015-06-30 2018-06-07 Omron Corporation Data flow control device and data flow control method
US20230327957A1 (en) * 2020-08-27 2023-10-12 Siemens Aktiengesellschaft Centralized management of data flow maps for distributed edge node deployment

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6390692B2 (ja) * 2016-12-15 2018-09-19 オムロン株式会社 データ配信システム、指示装置、データ配信装置、センサ管理装置、データ配信方法、およびプログラム
JP7175731B2 (ja) 2018-12-06 2022-11-21 エヌ・ティ・ティ・コミュニケーションズ株式会社 ストレージ管理装置、方法およびプログラム
JP7150584B2 (ja) 2018-12-06 2022-10-11 エヌ・ティ・ティ・コミュニケーションズ株式会社 エッジサーバとそのプログラム
JP7150585B2 (ja) 2018-12-06 2022-10-11 エヌ・ティ・ティ・コミュニケーションズ株式会社 データ検索装置とそのデータ検索方法およびプログラム、エッジサーバとそのプログラム
KR102310391B1 (ko) 2018-12-27 2021-10-07 미쓰비시덴키 가부시키가이샤 에지 시스템, 정보 처리 방법 및 기록 매체에 저장된 정보 처리 프로그램
DE112019007213B4 (de) 2019-05-20 2023-05-04 Mitsubishi Electric Corporation Ontologieerzeugungssystem, Ontologieerzeugungsverfahren und Ontologieerzeugungsprogramm
US11592809B2 (en) * 2019-11-29 2023-02-28 Rohde & Schwarz Gmbh & Co. Kg System and method for measurement data management in a distributed environment
US11321341B2 (en) * 2020-05-15 2022-05-03 National Instruments Corporation Dynamic metadata extraction workflow for measurement data storage
WO2022157873A1 (ja) * 2021-01-21 2022-07-28 三菱電機株式会社 情報処理装置、情報処理方法及び情報処理プログラム

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH103480A (ja) * 1996-06-17 1998-01-06 Nec Corp 文書検索方法および装置
JPH1145274A (ja) * 1997-07-28 1999-02-16 Just Syst Corp 単語間の共起性を用いたキーワードの拡張方法およびその方法の各工程をコンピュータに実行させるためのプログラムを記録したコンピュータ読み取り可能な記録媒体
JP2005202485A (ja) * 2004-01-13 2005-07-28 Toshiba Corp 映像提示装置
JP5445722B1 (ja) 2012-09-12 2014-03-19 オムロン株式会社 データフロー制御指令発生装置およびセンサ管理装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3037210B2 (ja) 1997-06-12 2000-04-24 甲府日本電気株式会社 スイッチング電源制御方式
US20160224637A1 (en) * 2013-11-25 2016-08-04 Ut Battelle, Llc Processing associations in knowledge graphs

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH103480A (ja) * 1996-06-17 1998-01-06 Nec Corp 文書検索方法および装置
JPH1145274A (ja) * 1997-07-28 1999-02-16 Just Syst Corp 単語間の共起性を用いたキーワードの拡張方法およびその方法の各工程をコンピュータに実行させるためのプログラムを記録したコンピュータ読み取り可能な記録媒体
JP2005202485A (ja) * 2004-01-13 2005-07-28 Toshiba Corp 映像提示装置
JP5445722B1 (ja) 2012-09-12 2014-03-19 オムロン株式会社 データフロー制御指令発生装置およびセンサ管理装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3540613A4

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180157692A1 (en) * 2015-06-30 2018-06-07 Omron Corporation Data flow control device and data flow control method
US11748326B2 (en) * 2015-06-30 2023-09-05 Omron Corporation Data flow control device and data flow control method
US20230327957A1 (en) * 2020-08-27 2023-10-12 Siemens Aktiengesellschaft Centralized management of data flow maps for distributed edge node deployment
US12047249B2 (en) * 2020-08-27 2024-07-23 Siemens Industry Software Inc. Centralized management of data flow maps for distributed edge node deployment

Also Published As

Publication number Publication date
US20190266165A1 (en) 2019-08-29
US10922325B2 (en) 2021-02-16
JP2018081377A (ja) 2018-05-24
EP3540613A4 (en) 2019-12-25
EP3540613A1 (en) 2019-09-18
EP3540613B1 (en) 2024-11-13
JP6406335B2 (ja) 2018-10-17

Similar Documents

Publication Publication Date Title
JP6406335B2 (ja) マッチング装置、マッチング方法及びプログラム
US20190278787A1 (en) Personalized content delivery system
RU2720905C2 (ru) Способ и система для расширения поисковых запросов с целью ранжирования результатов поиска
EP2864981B1 (en) Dynamic language model
US11410063B2 (en) Self-intelligent improvement in predictive data models
JP5795101B2 (ja) リアルタイムフィードバックを伴う情報検索システム
JP6782802B2 (ja) 計算機システム、計算機システムにより実行される方法及びプログラム
US9965522B2 (en) Mobile application search ranking
US20130346396A1 (en) Automatically updating a query
US20180308026A1 (en) Identifying risk patterns in a multi-level network structure
US20200279182A1 (en) Method and system for automatically producing plain-text explanation of machine learning models
JPWO2011142225A1 (ja) 特徴点検出システム、特徴点検出方法、及びプログラム
US11537674B2 (en) Method of and system for generating search query completion suggestion on search engine
JP2018084983A (ja) コンテンツマッチングシステム
US20160055203A1 (en) Method for record selection to avoid negatively impacting latency
US20190197050A1 (en) Search support system, search support method, and search support program
US20230342654A1 (en) Variable-output-space prediction machine learning models using contextual input embeddings
JP7354072B2 (ja) 情報処理装置、情報処理方法、及び情報処理プログラム
JP6168963B2 (ja) 情報検索装置、情報検索方法、およびプログラム
JP5741448B2 (ja) 情報推定装置、情報推定方法、及びプログラム
JP2019057034A (ja) 学習装置、生成装置、学習方法、生成方法、学習プログラム、生成プログラム、及びモデル
JP7375096B2 (ja) 分散表現生成システム、分散表現生成方法及び分散表現生成プログラム
JP4562533B2 (ja) 管制指示生成装置及びプログラム
JP2007025798A (ja) 言語知識構築支援プログラム及び方法
US20250356295A1 (en) Supply chain management apparatus, supply chain management method, and supply chain management system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17869161

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2017869161

Country of ref document: EP