Analysis support server and analysis support method

Abstract

This analysis support server includes: an acquiring unit that acquires time-series data including data at a plurality of time points, measured by a sensor device specified by a user; a submitting unit that submits an analysis request including the time-series data and a question relating to the time-series data to an analysis server; and a presenting unit that presents an answer to the question from the analysis server to the user.

Description

Analysis support server and analysis support method

[0001] The present invention relates to an analysis support server and an analysis support method.

[0002] In recent years, chatbots using large-scale language models have been put to practical use. Chatbots can also be used for data analysis. Patent Document 1 proposes a technique for analyzing stock price data, which is time-series data, using a language model.

[0003] Japanese Patent Application Laid-Open No. 2019-46158

[0004] It is conceivable that a user may designate a specific device and analyze the time-series data. An object of some aspects of the present invention is to provide a technique for analyzing time-series data related to a device in a simple manner.

[0005] According to some embodiments, there is provided an analysis support server including: a first acquisition means for acquiring time series data related to a device designated by a user; a second acquisition means for acquiring a question related to the time series data from the user; an input means for inputting an analysis request including the time series data and the question to an analysis server; and a presentation means for presenting an answer to the question to the user based on an analysis result by the analysis server.

[0006] Some embodiments provide a simplified method for analyzing time series data relating to a device. Other features and advantages of the present invention will become apparent from the following description taken in conjunction with the accompanying drawings, in which like reference numerals refer to like or similar components throughout the drawings.

[0007] The accompanying drawings are incorporated into and constitute a part of the specification, illustrate embodiments of the present invention, and together with the description are used to explain the principles of the present invention. A block diagram illustrating an example configuration of an analysis system according to some embodiments. A block diagram illustrating an example hardware configuration of components of an analysis system according to some embodiments. A block diagram illustrating an example hardware configuration of components of an analysis system according to some embodiments. A flow diagram illustrating an example operation of a management server according to some embodiments. A flow diagram illustrating an example operation of an analysis support server according to some embodiments. A sequence diagram illustrating an example operation of an analysis support server according to some embodiments.

[0008] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention as claimed, and not all combinations of features described in the embodiments are necessarily essential to the invention. Two or more of the features described in the embodiments may be combined in any desired manner. Furthermore, the same reference numerals are used to designate identical or similar components, and redundant descriptions will be omitted.

[0009] An example configuration of an analysis system 100 according to some embodiments will be described with reference to Fig. 1. The analysis system 100 may include a management server 101, an analysis support server 104, and an analysis server 106.

[0010] The management server 101 manages one or more sensor devices 102. Since the management server 101 typically manages multiple sensor devices 102, multiple sensor devices 102 are shown in FIG. 1 . For example, the management server 101 receives sensor data from each of the one or more sensor devices 102 and stores time-series data 111 composed of the sensor data in the storage 103. The sensor devices 102 may be devices having sensors. The sensor devices 102 may be capable of transmitting sensor data to the management server 101 via a network. The network between the sensor devices 102 and the management server 101 may be a cellular network, a local area network, a wide area network (e.g., the Internet), or any combination thereof. The sensor devices 102 may be communication devices, particularly Internet of Things (IoT) devices. The communication devices may be devices connectable to a network. The IoT devices may be devices connectable to the Internet. The embodiments described below are applicable not only to sensor devices 102 (i.e., devices having sensors) but also to devices without sensors. A device without a sensor may periodically transmit numerical data acquired by calculation to the management server 101. A device without a sensor may also be a communication device, particularly an IoT device.

[0011] The analysis server 106 provides a service that answers input questions. In particular, the analysis server 106 may answer questions by analyzing input data. The analysis server 106 may generate answers using, for example, a large-scale language model trained by machine learning. The analysis server 106 may also provide a chatbot such as ChatGPT by OpenAI. The chatbot may provide answers in natural language to questions posed in natural language.

[0012] The analysis support server 104 provides a service that supports the analysis of the time-series data 111 using the analysis server 106. Hereinafter, such a service will be referred to as an analysis support service. For example, the analysis support server 104 may provide the analysis support service to a user via the client terminal 105. The analysis support server 104 is capable of communicating with each of the management server 101, the client terminal 105, the analysis server 106, and the storage 103. Any network may be used to connect these devices.

[0013] The storage 103 stores data used by the analysis system 100. The storage 103 may be, for example, a database. The storage 103 may store time-series data 111 acquired from one or more sensor devices 102, a candidate question list 112, and past cases 113. The candidate question list 112 and past cases 113 will be described later. In FIG. 1 , the time-series data 111, the candidate question list 112, and the past cases 113 are stored in a single storage 103. Alternatively, these data may be stored in a distributed manner across multiple storages. At least a portion of the data stored in the storage 103 may be stored in any one of the servers of the analysis system 100.

[0014] The management server 101, the analysis support server 104, and the analysis server 106 may all be configured by devices in an on-premise environment or by computing resources in a cloud computing environment. The management server 101 and the analysis support server 104 may be configured by separate devices or may be configured by the same device.

[0015] 2A , an example of the hardware configuration of a computer 200 according to some embodiments will be described. The computer 200 may be used as any of the management server 101, the analysis support server 104, the client terminal 105, and the analysis server 106 described above.

[0016] The computer 200 may include the hardware devices shown in Fig. 2A. The processor 201 controls the overall operation of the computer 200. The processor 201 may be configured, for example, by a central processing unit (CPU), a graphics processing unit (GPU), or a combination of these. The processor 201 may be a single processor or a collection of multiple processors connected to each other so that they can communicate with each other.

[0017] The memory 202 stores programs and data used in the processing of the computer 200. The memory 202 may be configured, for example, by a combination of random access memory (RAM) and read-only memory (ROM).

[0018] The input device 203 is a device for obtaining instructions from a user of the computer 200. The input device 203 may be configured, for example, by a combination of one or more of a keyboard, buttons, a touchpad, and a microphone. The display device 204 is a device for visually presenting information to the user of the computer 200. The display device 204 may be, for example, a dot-matrix display such as a liquid crystal display. The computer 200 may have a device (e.g., a touch screen) in which the input device 203 and the display device 204 are integrated. The input device 203 and the display device 204 may be external to the computer. In this case, the computer 200 may have an interface for communicating with the external input device 203 and display device 204.

[0019] The communication device 205 is a device for communicating with devices external to the computer 200. When the computer 200 performs wired communication, the communication device 205 may be a network interface card (NIC) having a connector for connecting a cable. When the computer 200 performs wireless communication, the communication device 205 may be a wireless communication module including an antenna and a baseband processing circuit.

[0020] The secondary storage device 206 is a device for non-volatilely storing programs and data used in the processing of the computer 200. The secondary storage device 206 is configured by, for example, a hard disk drive (HDD) or a solid state drive (SSD).

[0021] In the following description, a component (e.g., processor 201) of computer 200 that functions as a specific device (e.g., management server 101) will be referred to as a component of this specific device (e.g., processor 201 of management server 101).

[0022] With reference to FIG. 2B , an example hardware configuration of the sensor device 102 according to some embodiments will be described. The sensor device 102 may have the hardware devices shown in FIG. 2B . The processor 211 to the communication device 215 may have similar functions to the processor 201 to the communication device 205. The sensor device 102 may have a simpler configuration than a general computer 200 (e.g., a smartphone or a personal computer). For example, the input device 213 may be a hardware button. The display device 214 may be an indicator. The sensor device 102 may not include at least one of the input device 213 and the display device 214.

[0023] The sensor 216 is a device for measuring specific data. The sensor 216 may be any type of sensor, such as a temperature sensor, a humidity sensor, an acceleration sensor, an illuminance sensor, a vibration sensor, a positioning sensor, a pressure sensor, an acoustic sensor, or an image sensor. The sensor device 102 may include multiple types of sensors 216. For example, if the sensor device 102 is a thermo-hygrometer, the sensor device 102 may include a temperature sensor and a humidity sensor as the sensors 216. The sensor device 102 may be powered by a battery (not shown) or an external power source.

[0024] An example of the operation of the management server 101 will be described with reference to FIG. 3 . The operation of FIG. 3 may be performed by the processor 201 of the management server 101 executing a program read into the memory 202 of the management server 101. Alternatively, some or all of the steps of the operation of FIG. 3 may be performed by a dedicated integrated circuit such as an application-specific integrated circuit (ASIC). The same applies to the operations of the analysis support server 104 and the analysis server 106 described below. The operation of FIG. 3 may be started when an administrator of the management server 101 starts up the management server 101. In the description of FIG. 3 , the management server 101 acquires sensor data from the sensor device 102. Alternatively, the management server 101 may acquire data from a device that does not have a sensor.

[0025] In S301, the management server 101 determines whether sensor data has been received from any of the multiple managed sensor devices 102. If the management server 101 determines that sensor data has been received ("YES" in S301), it transitions the process to S302, and otherwise ("NO" in S301), it repeats S301. The sensor device 102 may transmit sensor data to the management server 101 every time it measures data, or may transmit sensor data from multiple points in time together to the management server 101.

[0026] In S302, the management server 101 identifies which of the multiple sensor devices 102 to be managed is the sensor device 102 that transmitted the sensor data received in S301. The identified sensor device 102 is referred to as a target device. The management server 101 may identify the target device by an identifier added to the sensor data. This identifier may be, for example, an identifier used for communication (e.g., an Internet Subscriber Identity Number (IMSI)), or an identifier uniquely assigned by the management server 101 to each sensor device 102. Alternatively or additionally, the management server 101 may identify the target device based on the source address of the sensor data.

[0027] In S303, the management server 101 updates the time series data 111 stored in the storage 103. The management server 101 manages the time series data 111 separately for each of the multiple sensor devices 102 that it manages. If the time series data 111 for the target device is not stored in the storage 103, the management server 101 stores the sensor data received in S301 in the storage 103 as the time series data 111 for the target device. If the time series data 111 for the target device is stored in the storage 103, the management server 101 adds the sensor data received in S301 to the time series data 111 for the target device.

[0028] The management server 101 may assign a timestamp to the sensor data at each point in time when storing the sensor data in the storage 103. This timestamp may represent the date and time when the sensor data was received in S301, or may represent the date and time when the processing of S303 was executed. If a timestamp has already been assigned to the sensor data received in S301, the management server 101 may use this timestamp as is, or may update this timestamp.

[0029] Thereafter, the management server 101 returns the process to S301 and waits for the reception of new sensor data. Through the above operations, the storage 103 stores separate time-series data 111 for each of the multiple sensor devices 102 to be managed.

[0030] An example of the operation of the analysis support server 104 will be described with reference to FIG. 4 . The operation of FIG. 4 may be started in response to receiving an instruction from a user to the analysis support server 104 to analyze time-series data 111 related to one or more specific sensor devices 102. The specific sensor device 102 may be designated by the user using, for example, the identifier of the sensor device 102. The sensor device 102 designated by the user is referred to as a designated device. The sensor device 102 may be designated using an identifier used for communication (e.g., an Internet Subscriber Identity Number (IMSI)) or an identifier uniquely assigned to each sensor device 102 by the management server 101. The designated device may be one sensor device 102 or multiple sensor devices 102. When one sensor device 102 is designated, the time-series data 111 related to this single sensor device 102 is analyzed. When multiple sensor devices 102 are designated, the time-series data 111 related to these multiple sensor devices 102 are analyzed collectively. The designated device may be a sensor device 102 owned by the user who instructed the analysis, or may be a sensor device 102 not owned by this user. The time-series data relating to the sensor device 102 not owned by the user who instructed the analysis may be so-called open data.

[0031] The analysis support server 104 may function as a web server that provides a graphical user interface to the user through the client terminal 105. Specifically, the analysis support server 104 may present an operation screen to the user through the client terminal 105 and acquire instructions for this operation screen.

[0032] In S401, the analysis support server 104 acquires the time series data 111 of the designated device stored in the storage 103. The analysis support server 104 may read the time series data 111 of the designated device directly from the storage 103, or may read it via the management server 101. The analysis support server 104 may or may not present the acquired time series data 111 to the user.

[0033] In S402, the analysis support server 104 acquires a question for the analysis server 106 from the user. This question may be a question selected by the user from one or more candidate questions presented by the analysis support server 104, or may be a question arbitrarily entered by the user. For example, the question may be a sentence expressed in natural language, such as "Please explain the data." The question may be an expression of what the user wants to know. The question may be expressed as a declarative sentence or an interrogative sentence.

[0034] The candidate questions may be questions that can be applied to the time-series data 111 of any type of sensor device 102 (e.g., "Please explain the data"). Such questions are referred to as generic questions. Generic questions may be determined in advance by an administrator of the analysis support server 104, and may be included in the candidate question list 112 and stored in the storage 103. The types of sensor devices 102 may be classified according to the types of sensors 216 that the sensor devices 102 have.

[0035] A candidate question may be a question applicable to the time-series data 111 of a specific type of sensor device 102. For example, if the sensor device 102 is a thermo-hygrometer, a question such as "Is there any specific correlation between changes in temperature and changes in humidity?" may be asked. Such a question is referred to as a device-specific question. The device-specific question may be determined in advance by an administrator of the analysis support server 104 and stored in the storage 103 as part of the candidate question list 112. Alternatively or additionally, the device-specific question may be generated by the analysis server 106, as described below.

[0036] In S403, the analysis support server 104 submits an analysis request including the time-series data 111 acquired in S401 and the question acquired in S402 to the analysis server 106. This analysis request may be, for example, the following sentence: "Please answer based on the following CSV data: imsi, timestamp, temp, humi 001011234567890, 1685931995638, 23.4, 56.7 001011234567890, 1685932295638, 23.5, 56.8 (Omitted) 001011234567890, 1685942295638, 24.6, 55.5 Please explain the data." In addition to the time series data and the initial question, the analysis request may include an instruction to perform analysis on the time series data (in the above example, "Please answer based on the following CSV data") This instruction may specify the data format of the time series data (in the above example, CSV (Comma Separated Value) format). The analysis support server 104 may automatically generate an instruction statement based on the data format of the time-series data. Alternatively, the instruction statement may simply be a statement such as "Please respond based on the following data" without specifying the data format of the time-series data.

[0037] The first analysis request may be made in a new context. The context may refer to a series of data input by the analysis support server 104 to the analysis server 106. In other words, making an analysis request in a new context may mean making a new analysis request regardless of data previously input.

[0038] The analysis support server 104 may include all of the time-series data 111 acquired in S401 in the analysis request, or may include only a portion of the time-series data 111 in the analysis request to reduce the amount of data. For example, the analysis support server 104 may include only a specific number or a specific amount of data from the time-series data 111 acquired in S401 in the analysis request. Alternatively or additionally, the analysis support server 104 may include only a portion of the time-series data 111 acquired in S401 for a specific time period in the analysis request. Alternatively or additionally, the analysis support server 104 may include only data sampled at predetermined intervals from the time-series data 111 acquired in S401 in the analysis request. Alternatively or additionally, the analysis support server 104 may include only data of predetermined items (e.g., only timestamps and temperature data) from the time-series data 111 acquired in S401 in the analysis request. The criteria for extracting sensor data may be set in advance by an administrator of the analysis support server 104 or may be specified by a user. When an analysis request includes time-series data related to multiple sensor devices 102, the time-series data included in the analysis request may be sorted based on the sensor device, may be sorted based on the timestamp, or may be arranged in any order.

[0039] The analysis server 106 performs an analysis in response to the analysis request submitted by the analysis support server 104, and returns the analysis results to the analysis support server 104. The analysis support server 104 receives the analysis results. In S404, the analysis support server 104 presents an answer to the user based on the analysis results received from the analysis server 106. The analysis support server 104 may present the analysis results to the user as they are, or may process the analysis results before presenting them to the user. For example, the analysis support server 104 may highlight parts of the analysis results in accordance with predetermined rules (for example, it may highlight keywords such as "important," "necessary," and "attention").

[0040] Before or after presenting the answer in S404, the analysis support server 104 may store the question submitted to the analysis server 106 and the device information of the specified device in the storage 103 as past cases 113. The device information may include, for example, the type of the specified device. The past cases 113 may be used to generate candidate questions, as described below.

[0041] In S405, the analysis support server 104 determines whether the user has instructed the user to ask an additional question. If the analysis support server 104 determines that the user has instructed the user to ask an additional question (YES in S405), the process proceeds to S403. Otherwise (NO in S405), the process ends.

[0042] When an additional question is instructed, the analysis support server 104 executes S403 again and submits an analysis request including the additional question to the analysis server 106. This analysis request may be, for example, the following sentence: "Please explain why you identified noteworthy data." This analysis request may be made in the same context as the previous question. If the analysis server 106 has a function for managing context, the analysis request including the additional question may not include the time-series data 111 acquired in S401. On the other hand, if the analysis server 106 does not have a function for managing context, the analysis support server 104 may send to the analysis server 106 a new analysis request that includes an additional question in addition to a previously made analysis request.

[0043] If no additional questions are instructed by the user, the analysis support server 104 ends the processing. In the above example, each time an analysis request is submitted to the analysis server 106, the analysis support server 104 stores the content of the request as a past case 113. Alternatively, the analysis support server 104 may collectively store the content of the analysis request as a past case 113 after a series of questions from the user has been completed.

[0044] The past case 113 may include other information in addition to or instead of the content input to the analysis server 106 as described above. For example, the past case 113 may include text describing past cases conducted using the sensor device 102 or other devices. Such text may have titles such as "Visualizing room temperature data and power usage and remotely controlling air conditioners to provide an energy-saving service," "Developing a tire pressure and temperature management service using IoT: A long-established tire manufacturer embarks on a new business venture," "Digitizing the reasons for production line shutdowns at a long-established factory and improving operations through data analysis," or "Automating safety management of distribution boards within manufacturing plants using IoT: Temperature data collection makes it possible to detect abnormalities in distribution boards," and may be published as a webpage. The text of such past case 113 may be written by a person and stored in the storage 103. Alternatively, the text of the past case 113 may be automatically generated by the analysis server 106 or another server.

[0045] An example of the operation in which the analysis support server 104 presents candidate questions to a user will be described with reference to FIG. 5 . In the method of FIG. 5 , the analysis server 106 generates two instances for managing two independent contexts. One instance manages a context for question generation, and therefore this instance is referred to as a question generation instance 501. The other instance manages a context for data analysis, and therefore this instance is referred to as an analysis instance 502. In the example of FIG. 5 , the question generation instance 501 and the analysis instance 502 are included in one analysis server 106. Alternatively, the question generation instance 501 and the analysis instance 502 may be included in two separate analysis servers 106. If the analysis server 106 does not have the function of managing contexts, the analysis support server 104 may execute question generation and data analysis in separate contexts by modifying the content of a previous request to be included in a new request.

[0046] In S501, the analytical support server 104 acquires an instruction to start analytical support from the user 500. As described above with reference to Fig. 4, the start instruction includes an identifier for identifying the designated device.

[0047] In S502, the analysis support server 104 acquires the time-series data 111 of the specified device in the same manner as in S401. In S503, the analysis support server 104 requests the question generation instance 501 to analyze the time-series data 111 acquired in S502. This analysis is performed to generate a basis for generating candidate questions. This analysis request may be, for example, the following sentence: "Please tell us which three data series have the largest fluctuations among the following data. Also, what do these data series represent? --- __time,__iso8601Time,bat,humi,lat,lon,rs,temp,type,x,y,z "2023-06-21 07:04:34","2023-06-21T07:04:34.048+09:00","3","53.7","35.701044","139.673169","4","28.1","0","0","-192","1024" "2023-06-21 07:03:39","2023-06-21T07:03:39.008+09:00","3","53.4","35.701042","139.67317","4","28.1","0","0","-192","1024" "2023-06-21 07:02:34","2023-06-21T07:02:34.053+09:00","3","53.4","35.70104","139.673171","4","27","0","0","-192","1024" "2023-06-21 07:01:33","2023-06-21T07:01:33.962+09:00","3","53.3","35.701039","139.67317","4","28.1","0","0","-192","1024" "2023-06-21 07:00:33","2023-06-21T07:00:33.978+09:00","3","53.7","35.701039","139.67317","4","28.2","0","0","-192","1024": (Omitted below) ---"

[0048] In S504, the question generation instance 501 executes the analysis. In S505, the question generation instance 501 responds with the analysis results to the analysis support server 104. For example, the following sentence is returned: "The following three data series have large changes. - humi: The humidity value has changed significantly from 53.9 to 47.2. - temp: The temperature value has changed significantly from 28.4 to 27. - x: The acceleration value in the x-axis direction has changed significantly from 64 to -192. This data appears to be sensor data that records the date and time, remaining battery level, humidity, latitude and longitude, temperature, acceleration, etc."

[0049] In S506, in the same context as S503, the analysis support server 104 requests the question generation instance 501 to propose examples that can be used as references for generating candidate questions for the time-series data 111 acquired in S502. This request may be, for example, the following sentence: "Please consider what examples you should use as references to analyze a data series with the following characteristics. For example, if there is a large change in latitude and longitude, it would be a good idea to refer to examples that use GPS. --- The following three data series have large changes. - humi: The humidity value changed significantly from 53.9 to 47.2. - temp: The temperature value changed significantly from 28.4 to 27. - x: The acceleration value in the x-axis direction changed significantly from 64 to -192. This data appears to be sensor data that records the date and time, remaining battery level, humidity, latitude and longitude, temperature, acceleration, etc. ---" The analysis support server 104 may generate the sentence in S506 by adding an instruction statement preset by an administrator to the analysis results received in S505.

[0050] In S507, the question generation instance 501 generates a case suggestion. In S508, the question generation instance 501 responds with the case suggestion to the analysis support server 104. For example, the following sentence is responded: "- humi: When humidity changes suddenly, it may be affected by the indoor temperature or the outdoor humidity. For example, humidity changes when the temperature setting of an air conditioner is changed, or humidity changes after it rains. It may be helpful to refer to these examples. - temp: When temperature changes suddenly, it may be affected by the indoor temperature or the outdoor temperature. For example, temperature changes when the temperature setting of an air conditioner is changed, or changes due to temperature differences between daytime and nighttime. It may be helpful to refer to these examples. - x: When acceleration changes suddenly, it may be affected by movement of the sensor. For example, acceleration changes when the sensor is shaken or dropped. It may be helpful to refer to these examples."

[0051] In S509, the analysis support server 104 identifies the case proposed in S508 from the past cases 113. For example, the analysis support server 104 may calculate the similarity between each of the multiple sentences included in the past cases 113 and the sentence received in S508, and identify the case where this similarity exceeds a threshold as the recommended case. The similarity between sentences may be calculated using existing technology.

[0052] In S510, in the same context as S506, the analysis support server 104 requests the question generation instance 501 to generate candidate questions for the time-series data 111 acquired in S502. This request may be, for example, the following sentence: "[Text of the identified reference case] --- - humi: When humidity changes suddenly, it may be affected by the indoor temperature or outdoor humidity. For example, humidity changes when the temperature setting on an air conditioner is changed, or humidity changes after it rains. You may refer to these examples. - temp: When temperature changes suddenly, it may be affected by the indoor temperature or outdoor temperature. For example, temperature changes when the temperature setting on an air conditioner is changed, or changes due to temperature differences between daytime and nighttime. You may refer to these examples. - x: When acceleration changes suddenly, it may be affected by sensor movement. For example, acceleration changes when the sensor is shaken or dropped. You may refer to these examples. --- Focusing on the data above, prepare five questions to confirm the data. If the reference material states that the same type of data is being used, please ask questions related to that. - Please state your questions in bullet points. - Answer in Japanese. - Ask questions to gain insights from changes in the data. - Do not ask about the cause of the change in the data. - Do not ask about measures to address changes in the data. - Do not ask about case studies. - Do not ask about sensor specifications. - Do not ask about where the sensor is installed. - Do not ask about data that the sensor cannot obtain. - Please use your imagination to fill in any gaps in the context. Example questions: Q: At what time did the sudden change in the data occur? Q: Are there any cyclical changes in the data? Q: Are there any cases where the temperature changes suddenly or shows extreme values? The text of the case identified in S509 may be used as is in the [Text of the identified reference case] included in the above text.If multiple sentences are identified in S509, the "sentence of the identified reference case" included in the above sentence may contain multiple sentences. If the analysis server 106 has a function for analyzing web pages, the "sentence of the identified reference case" included in the above sentence may contain the URL of the web page in which the reference case is described.

[0053] In S511, the question generation instance 501 generates candidate questions. In S512, the question generation instance 501 responds with the candidate questions to the analysis support server 104. For example, the following sentence is returned: "Q: Are there any distinctive patterns or trends in temperature or humidity values ​​depending on the day of the week or time of day? Specifically, for example, is there a pattern where the temperature is higher every Monday morning compared to other days of the week or time of day? Q: Are there any distinctive patterns in the amount of temperature change? For example, are there regularly periods or time periods where the amount of temperature change is large and periods or time periods where the amount of temperature change is small? Q: Is there any special correlation between changes in humidity and changes in temperature? Q: Is there any periodicity in the time periods where the amount of data change is distinctive? In other words, are there any periods where a specific pattern appears regularly? Q: What are the maximum and minimum changes in temperature and humidity? Are there any specific conditions or patterns under which these data changes are observed?"

[0054] In S513, the analysis support server 104 presents the candidate questions responded to by the question generation instance 501 to the user 500. The analysis support server 104 may present to the user, in addition to the candidate questions responded to by the question generation instance 501, candidate questions included in the candidate question list 112 stored in the storage 103. If the quality of the candidate questions responded to by the question generation instance 501 is low or if a response from the question generation instance 501 is not obtained within a threshold time, the analysis support server 104 may present to the user only the candidate questions included in the candidate question list 112 stored in the storage 103. The analysis support server 104 may request the question generation instance 501 to evaluate the quality of the candidate questions responded to by the question generation instance 501. The analysis support server 104 may present the candidate questions included in the candidate question list 112 stored in the storage 103 and the candidate questions responded to by the question generation instance 501 in a manner that allows them to be distinguished.

[0055] In S514, the analysis support server 104 acquires a question from the user 500, similar to S402. In S515, the analysis support server 104 requests an analysis from the analysis instance 502, similar to S403. This request is made in a new context. Furthermore, this request is made in a context different from the context for generating candidate questions made in S503 to S512 above. In S516, the analysis server 106 executes the analysis. In S517, the analysis server 106 responds with the analysis results. In S518, the analysis support server 104 presents an answer to the user 500 based on the analysis results, similar to S404. As described above, the analysis support server 104 may present the analysis results to the user as they are, or may process the analysis results before presenting them to the user.

[0056] In S519, the analysis support server 104 requests that the context stored in the analysis instance 502 be copied to the question generation instance 501. In S520, the context stored in the analysis instance 502 is copied to the question generation instance 501. Thereafter, in S521, the same processing as in S506 to S518 is executed.

[0057] By copying the context of the analysis instance 502 to the question generation instance 501, candidate questions are generated based on the analysis performed in S516. If the analysis server 106 does not have a function for managing contexts, the analysis support server 104 may make a question generation request in the same context as the analysis request by including the contents of the analysis request submitted in S515 in the question generation request in S521. In this case, S515 and S516 are not executed.

[0058] Thereafter, the analysis support server 104 repeats steps S519 to S521 each time it receives an additional question from the user 500. The analysis request for the additional question is made in an existing context stored in the analysis instance 502. Furthermore, this request is made in a context different from the context for generating candidate questions made in step S521 above.

[0059] In S522, the analysis support server 104 acquires an end instruction from the user 500. The end instruction may be an explicit end instruction (such as pressing an end button) or an implicit end instruction (such as closing the browser or the passage of a predetermined time).

[0060] 5 , the analysis support server 104 presents candidate questions to the user for each question from the user. Alternatively, the analysis support server 104 may present candidate questions to the user only before the first question from the user, and may not present candidate questions to the user thereafter. Alternatively, the analysis support server 104 may not present candidate questions to the user before the first question from the user, but may present candidate questions to the user before the second or subsequent questions.

[0061] The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the gist of the invention.

[0062] This application claims priority based on Japanese Patent Application No. 2023-110214, filed on July 4, 2023, the entire contents of which are incorporated herein by reference.

Claims

1. An analysis support server comprising: a first acquisition means for acquiring time series data relating to a device specified by a user; a second acquisition means for acquiring a question relating to the time series data from the user; an input means for inputting an analysis request including the time series data and the question to an analysis server; and a presentation means for presenting an answer to the question to the user based on an analysis result by the analysis server.

2. The analysis support server according to claim 1, wherein the analysis request includes a timestamp associated with data at each point in time of the time series data.

3. The analysis support server according to claim 1 or 2, wherein the presenting means further presents candidate questions to the user.

4. The analysis support server according to claim 3, wherein the input means further inputs the time series data and a question generation request to generate candidate questions regarding the time series data to the analysis server or another analysis server, and the presentation means presents the candidate questions generated by the analysis server or the another analysis server to the user.

5. The analysis support server according to claim 4, wherein the question generation request includes past cases based on suggestions by the analysis server or the other analysis server.

6. The analysis support server according to claim 4 or 5, wherein said input means inputs said analysis request and said question generation request in different contexts.

7. The analysis support server according to any one of claims 1 to 6, wherein the analysis request specifies a data format of the time series data.

8. The analysis support server according to any one of claims 1 to 7, wherein the device is equipped with a sensor.

9. The analysis support server according to any one of claims 1 to 8, wherein the second acquisition means further acquires a follow-up question to the answer from the user.

10. A program for causing a computer to function as the analysis support server according to any one of claims 1 to 9.

11. An analysis support method comprising: a first acquisition step in which a first acquisition means acquires time series data relating to a device specified by a user; a second acquisition step in which a second acquisition means acquires a question relating to the time series data from the user; an input step in which an input means inputs an analysis request including the time series data and the question to an analysis server; and a presentation step in which a presentation means presents an answer to the question to the user based on an analysis result by the analysis server.

Images