KR101916934B1 - Data searching apparatus - Google Patents

Abstract

A data search apparatus according to an embodiment of the present invention includes a memory for storing time series data; And a processor accessible to the memory, wherein the processor can assign an externally input comment at a part of or a part of the time series data, and classify the comment according to a classification tag included in the comment .

Description

DATA SEARCHING APPARATUS

The present invention relates to a data search apparatus.

Since data can be collected by anyone through web, smart phone, IoT sensor, etc., data sources are diversified and personalized. In order to support this, open-source data analysis algorithms and platform services have been developed. In addition, the algorithm can be applied without expert knowledge.

However, even if data and algorithms are prepared, it is not possible for anyone to easily utilize the data. It is possible to process data, search key information contained in the data, or perform data mining or machine learning algorithms ) Requires technical knowledge and experience, but not all of them have such knowledge and experience.

In the future, as well as expert knowledge of data or algorithms, experience knowledge of environment and circumstances in which data is generated, personal tendencies, and know-how about which algorithm to use with which parameters are applied to certain data Will become even more important.

In addition, the process of collecting large amounts of such data together with data is a very important factor in implementing artificial intelligence services.

Therefore, in order to make the best use of all of their data, everyone should be able to easily borrow the ability of the experienced person who possesses the knowledge of the data or the expert of the data analysis. Meanwhile, Through these processes, we need to be able to use our knowledge and experience to create opportunities for profit.

[Patent Document 1] Japanese Patent Application Laid-Open No. 10-2007-0108294 (Publication Date: November 09, 2007)

A data search apparatus according to an embodiment of the present invention is for searching for data of a pattern desired by a user from search time series data.

A data search apparatus according to an embodiment of the present invention is for classifying comments allocated in a matching section.

The data searching apparatus according to the embodiment of the present invention is for estimating the prices of comments, setting intervals, and classification tag lists.

The data search apparatus according to an embodiment of the present invention classifies newly inputted analysis target time series data according to classification tags.

The data search apparatus according to an embodiment of the present invention is for assigning a comment to a section or point selected by a user.

The data search apparatus according to an embodiment of the present invention is for estimating the prices of comments, selection periods, selection points, and classification tag lists.

The task of the present application is not limited to the above-mentioned problems, and another task which is not mentioned can be clearly understood by a person skilled in the art from the following description.

According to an aspect of the present invention, there is provided an information processing apparatus including: a memory for storing time series data; And a processor accessible to the memory, wherein the processor can assign an externally input comment at a part of or a part of the time series data, and classify the comment according to a classification tag included in the comment .

The processor may generate a comment list for the comment, associate the comment list with the classification tag, and generate a classification tag list for the classification tag.

Wherein the processor is configured to receive and assign a score for at least one of the classification tag and the comment from one or more user terminals and to calculate the number of citations of the comment when the comment is cited in another comment, The price of the comment can be calculated according to the number of quotations of the comment.

The processor may generate a data vector having the characteristics of the time series data to which the comment is assigned and apply another time series data to the machine learning model according to the data vector to classify the time series data according to the classification tag have.

The processor may apply the time series data to the machine learning model without comment assignment.

The data search apparatus according to the embodiment of the present invention can search for desired data by searching the search target time series data for a portion that matches or is similar to the pattern corresponding to the setting interval in a plurality of different time series data.

The data search apparatus according to the embodiment of the present invention can classify the comments allocated in the matching section according to the classification tag by allocating the comments including the classification tag.

The data searching apparatus according to the embodiment of the present invention can calculate the price according to the score or the number of citations according to the appropriateness of the comment, the setting period, and the classification tag list.

The data searching apparatus according to the embodiment of the present invention can classify newly inputted analysis target time series data according to classification tags through a machine learning model.

The data search apparatus according to the embodiment of the present invention can calculate the price through the score, the number of citations, and the number of comments according to the appropriateness of the comment, the selection period, the selection time, and the classification tag list.

The data searching apparatus according to the embodiment of the present invention can classify newly inputted analysis target time series data according to classification tags through a machine learning model.

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

1 shows a data searching apparatus according to an embodiment of the present invention.
2 shows an example of the first time series data, the second time series data, the first search target time series data, and the second search target time series data.
FIG. 3 shows an example of a comment assigned to a matching section.
4 shows an example of a classification tag list.
5 and 6 are diagrams illustrating a process of generating a data vector.
Figures 7-9 show through a machine learning model.
10 to 12 show comments assigned to the section selected by the user.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. It is to be understood, however, that the appended drawings illustrate the present invention in order to more easily explain the present invention, and the scope of the present invention is not limited thereto. You will know.

Also, the terms used in the present application are used only to describe certain embodiments and are not intended to limit the present invention. The singular expressions include plural expressions unless the context clearly dictates otherwise.

In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

1 shows a data searching apparatus according to an embodiment of the present invention. 1, a data search apparatus according to an embodiment of the present invention includes a memory 106 and a processor 104. [

The data searching apparatus according to an embodiment of the present invention may include a bus 102 or other communication mechanism for transferring information. Such a bus 102 or other communication mechanism may include a processor 104, a computer readable recording medium RM, a network interface 112 (e.g., a modem or Ethernet card), a display portion 114 A keyboard, a keypad, a virtual keyboard, a mouse, a trackball, a stylus, a touch sensing means, etc.), and / or subsystems.

The computer readable recording medium RM includes a memory 106 (e.g., RAM), a static storage 108 (e.g., ROM), a disk drive 110 (e.g., HDD, SSD, Disk, flash memory drive, etc.). At this time, the disk drive may be a non-transitory recording medium. The optical disc may be a CD, a DVD, or a Blu-ray disc, but is not limited thereto.

The data searching apparatus according to an embodiment of the present invention may include one or more disk drives 110. Also, as shown in FIG. 1, the disk drive 110 may be included in the housing 120 along with the processor 104, but may be remotely located to perform remote communication with the processor 104. And may include a database having one or more disk drives.

The recording medium RM may store an operating system, a driver, an application program, data and a database necessary for the operation of the data searching apparatus according to the embodiment of the present invention.

The display unit 114 may display an operation of the data search apparatus and a user interface according to an embodiment of the present invention.

The processor 104 may be, but is not limited to, a CPU, a microcontroller, a digital signal processor (DSP), and the like, and controls the operation of the data search apparatus according to an embodiment of the present invention.

Processor 104 is connected to a recording medium RM to perform data searching, comment allocation, classification tag processing and machine learing (to be described later) by executing one or more sequences of instructions stored in a recording medium RM ) Can be performed.

These instructions may be read into the memory 106 from another computer readable medium such as the static storage 108 or the disk drive 110. [ In other embodiments, hard-wired circuitry may be used instead of or in combination with software instructions to implement the present disclosure.

The logic may be encoded in a computer readable recording medium (RM), which may refer to any medium that participates in providing instructions to processor 104 for execution. Such a recording medium RM may take many forms including, but not limited to, non-volatile recording media, volatile recording media.

The processor 104 may communicate with the hardware controller for the display unit 114 to display the operation of the data searching apparatus and the user interfacing operation on the display unit 114. [

In one embodiment, the computer readable recording medium RM may be non-volatile. Volatile recording media RM include optical or magnetic disks such as disk drive 110 and volatile recording media include a dynamic recording medium such as system memory 106 and a bus (not shown) 102 include coaxial cables, copper wire, and optical fibers.

In one example, transmission media may take the form of sound waves or light waves, such as those generated during radio and infrared data communications.

Some common forms of computer readable media (RM) are, for example, a floppy disk, a flexible disk, a hard disk, a magnetic tape, any other magnetic medium, a CD-ROM, , A paper tape, any other physical medium having a pattern of holes, a RAM, a PROM, an EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave, or any other medium adapted to be read by a computer.

In various embodiments of the present disclosure, execution of the instruction sequences for implementing the present disclosure may be performed by a data search apparatus according to an embodiment of the present invention. In various other embodiments of the present disclosure, other wired or wireless networks (e.g., LAN, WLAN, PTSN, and / or remote communications, including mobile and cellular telephone networks) A plurality of computing devices 100 coupled to a network may perform the instruction sequences for implementing the present disclosure in cooperation with each other.

The data searching apparatus according to the embodiment of the present invention transmits and receives commands including messages, data, information and one or more programs (i.e., application code) through the communication link 124 and the network interface 112 You may.

The network interface 112 may include separate or integrated antennas to enable transmission and reception over the communication link 124. The received program code may be executed by the process 104 when received and / or may be stored in the disk drive 110 or some other non-volatile store for execution.

Next, the operation of the data search apparatus according to the embodiment of the present invention will be described with reference to the drawings.

The memory 106 stores different first time series data and second time series data. The first time series data and the second time series data may include information on various data values over time.

For example, the first time series data and the second time series data may be information on a sensing value according to the time outputted by the sensor, or information on a stock price according to time of a specific company or a stock market.

The first time series data and the second time series data may be respectively output from different sensors for sensing the same factor or may be different from each other depending on different factors (for example, the temperature of the sea surface and the travel path of the typhoon, . ≪ / RTI >

The processor 104 may access the memory 106. Accordingly, the processor 104 can read the first time series data and the second time series data.

2 shows an example of the first time series data, the second time series data, the first search target time series data, and the second search target time series data. In FIG. 2, the horizontal axis corresponds to the time and the vertical axis corresponds to the value of each time series data.

The processor 104 derives the first matching data which is a part of the first search target time series data matched with the first pattern of the first time series data existing in the setting interval.

The processor 104 also derives second matching data that is part of the second search target time series data that is different from the first search target time series data that matches the second pattern of the second time series data existing in the setting interval.

That is, the processor 104 can search the first search target time series data and the second search target time series data that match the different time series data patterns existing in the same setting interval.

For this, the processor 104 searches a plurality of patterns of the set period for data of the same or similar pattern as a plurality of patterns while windowing the entire section of the plurality of search target time series data to obtain first matching data and second Matching data can be derived. Acceptable error rates can be provided with default values, and the default values are changeable.

For this purpose, the processor 104 may perform data sampling in the setup interval and search for data that is consistent with or similar to the order and data values of the sampled data.

The processor 104 may perform a search operation on the first search target time series data and the second search target time series data, but is not limited thereto. As shown in FIG. 2, the processor 104 may search for three or more different search target time series data The search operation can be performed.

As described above, the data search apparatus according to the embodiment of the present invention can search for data similar to or identical to the pattern of the setting interval among different time series data.

For example, when the first time series data to the third time series data are temperature, humidity, and crop growth amounts respectively with respect to time, the user can select a section in which the relation of temperature, humidity, When the data is input to the apparatus, the data search apparatus can search for a portion of the search target data that matches or is similar to the pattern of the temperature, humidity, and crop growth amount in the setting region. Accordingly, the user can easily find the section with high relevance.

Meanwhile, as shown in FIG. 2, the first search target time series data and the second search target time series data may be at least a part of the first time series data and the second time series data, respectively.

Alternatively, unlike FIG. 2, the first search target time series data and the second search target time series data may be different from the first time series data and the second time series data, respectively.

For example, the first time series data and the second time series data are the data output from the temperature sensor 1 and the humidity sensor 1, respectively, and the first search target time series data and the second search target time series data are the temperature sensor 2 and the humidity sensor 2 And may be output data.

Accordingly, the user can confirm whether or not a portion of the data of the temperature sensor 1 and the humidity sensor 1 that is similar to or coincides with the pattern of the setting section is in the data output from the temperature sensor 2 and the humidity sensor 2.

In the above description, the setting of the setting interval can be performed through a query.

The processor 104 may perform a search operation after smoothing a plurality of search target time series data to a certain degree with a normalization filter or a mean filter before the search.

Meanwhile, as shown in FIG. 3, the processor 104 may assign a comment input from the outside in a matching interval in which the first matching data and the second matching data exist. The user can input a comment of the matching section through his or her terminal or input unit 118. [ The comment may be a user's interpretation, comment, or memo about the matching interval, but is not limited thereto.

The terminal may be connected to the data search apparatus according to an embodiment of the present invention and may be a PC, a tablet, a smart phone, or a laptop, but is not limited thereto.

At this time, the comment may include a classification tag, and the processor 104 may classify the comment according to the classification tag included in the comment. For example, the user can input a comment saying 'If another matching interval occurs within # 1 minute after the first matching interval', the user is requested to pay attention.

At this time, the comment may include a classification tag such as # 1 minute. In an embodiment of the present invention, the classification tag may be a hash tag, but is not limited thereto.

As shown in FIG. 4, the processor 104 may generate a classification tag list sorted by classification tag, and may store the classification tag list in the recording medium RM.

That is, the processor 104 may generate a comment list for comments, associate it with a classification tag, and generate a classification tag list for the classification tag. A comment list can be generated for each classification tag, and can include comment-related information as well as comments.

That is, the comments 1 to 3 include the classification tag #ABCD, and the comments 4 to 6 may include the classification tag #WXYZ.

The comment related information may include the title of the comment, the ID of the comment creator, the creation time, the start and end positions of the matching interval to which the comment is assigned, and the maximum, minimum and average values of the data in the matching interval to which the comment is assigned But is not limited thereto.

In FIG. 4, comments 1 to 6 may be characters or symbols, but may be codes assigned to comments 1 to 6, respectively.

Meanwhile, the processor 104 may receive and allocate a score for at least one of a setting interval, a classification tag, and a comment from one or more user terminals.

That is, the specific user can assign a score for the appropriateness of the setting period, the classification tag, and the comment made by other users through the input unit 118 or the terminal. The score for the setting interval may be a score for the query that includes information about the setting interval.

The processor 104 may also calculate the number of times a comment is cited if the comment is cited in another comment. For example, when comment 3 is 'requesting attention according to comment 1 if another matching interval occurs within # 1 minute after the first matching interval', comment 3 may be quoted comment 1 once.

Through this, the processor 104 can calculate the price of the comment according to the number of citations and the number of citations of the comments. Accordingly, the data search apparatus according to the embodiment of the present invention can provide appropriate compensation to the user who created the corresponding comment.

In the above description, the price is calculated according to the number of quotations of the score and the comment, but the price is calculated according to the score of the appropriately set setting interval, and compensation can be provided to the user who sets the setting interval.

Meanwhile, the processor 104 may generate a data vector consisting of the first feature of the first matching data and the second characteristic of the second matching data. 5 and 6 are diagrams illustrating a process of generating a data vector.

The data vector generation process shown in FIG. 5 will be described and then the data vector generation process shown in FIG. 6 will be described.

In FIG. 5, Data # 1 to Data # 4 are first to fourth matching data that are different from each other. At this time, the data generation periods of the search target time series data may be different from each other.

For example, data of Data # 1 is generated every unit time (for example, 1 second), data of Data # 2 is generated every 2 seconds, and Data # 3 and Data # 4 are generated every 4 seconds and 6 seconds ≪ / RTI > In the above description, the unit time is 1 second, but the present invention is not limited thereto, and may be changed depending on the case.

In this case, the first to fourth features of the first to fourth matching data may be data values of each matching data. The data vectors may be made up of these data values.

Since the data generation cycle of the matching data may be different as described above, the first matching data has a data value in a specific unit time, but the second matching data has a data value It may not exist.

If the data value of the matching data does not exist, the data vector is not formed, so that the data value can be virtually generated according to the set method.

5, if there is no data value between the nth data value of each matching data and the n + Tth (n is a natural number of 1 or more, T period) data value, the processor 104, May virtually generate an nth data value to fill the nth data value and the n + mth data value.

That is, since the nth data value is # 1 and T is 2, the (n + 2) th (= n + T) data value is # 2 and the There is no data value.

Accordingly, the processor 104 can virtually fill the nth data value and the (n + 2) th (= n + T) data value with # 1 (= nth data value). Processor 104 may fill in the data values in the same manner with the remaining third and fourth matching data.

As shown in FIG. 5, there may be a period in which a data vector can not be generated. When generating the first data vector, in the case of Data # 2, there is no data at the present time, and it is necessary to fetch from the past data, and there may not exist the past data.

This phenomenon occurs even in the case of Data # 3 in the first and second data vectors, so that the first data vector and the second data vector can not be completely filled. Accordingly, the first data vector and the second data vector may not be generated.

Alternatively, an average value of the n-th data value and the n + T-th data value may be virtually generated to fill the empty data value, and the data value may be virtually generated through various methods.

Accordingly, a data vector can be generated for each unit time.

Meanwhile, the processor 104 may generate a data vector by sampling a part of the entire data of the matching interval. That is, the first characteristic and the second characteristic may be data values sampled from each of the first matching data and the second matching data at the same time point. A reliable data vector can be generated while reducing the amount of computation of the processor 104 by adjusting the sampling rate.

Next, a method of generating a data vector will be described with reference to FIG.

As shown in FIG. 6, the first matching data and the second matching data, which are different from each other, may be represented by straight lines connected to the display unit 114. At this time, the first characteristic of the first matching data and the second characteristic of the second matching data may include a slope of each of the segmented first matching data and the second matching data existing in the same section.

The processor 104 may perform segmentation on time series data of time series data, search time series data, or matching data, and may use a piecewise linear segmentation technique. Such a segmentation technique is not limited to the piecewise linear segmentation technique, and various segmentation techniques can be applied to the present invention.

Accordingly, the time series data, the search target time series data, or the matching data can be formed as a straight line segment.

As shown in FIG. 6, a data vector may be generated for each section. It should be determined whether the setting of the interval is based on the first matching data or the second matching data.

In the case of the embodiment of the present invention, the processor 104 may set intervals of matching data having a large number of segments. Accordingly, more data vectors can be generated when the second matching data is used as a reference.

As shown in FIG. 6, since the number of segments of the first matching data is larger than the number of segments of the second matching data, the reference matching data for setting the interval may be the first matching data.

The processor 104 may set the interval every time the slope of the segment constituting the first matching data changes, and may generate the data vector through the slope of the first matching data and the second matching data of each interval.

At this time, in the section A, the number of segments of the second matching data is larger than the number of segments of the first matching data which is the reference matching data. Accordingly, a plurality of segment slopes may exist in the section A, and the processor 104 may set a representative slope representing the plurality of segment slopes in the section A as the second characteristic.

In the embodiment of the present invention, the representative slope may be an average value of a plurality of segment slopes, but the present invention is not limited thereto and representative slopes can be set by various methods.

Also, the first and second features of the data vector may include data values at the boundary of the interval along with the slope. In Fig. 6, the data value of the black dot may be the data value at the section boundary.

If the generation of the data vector is not limited to the method shown in Figs. 5 and 6, the data vector can be generated by various methods.

The machine running model will be described in detail with reference to FIGS. 7 to 9 below.

As described above with reference to FIG. 5 and FIG. 6, the data vector is generated in the matching interval, and the comment including the classification tag is assigned to the matching interval, so that the data vector can be associated with the classification tag have.

The machine learning model of FIG. 7 uses decision tree learning. That is, the relationship between Data # 1, Data # 2, and Data # 3 of the data vector associated with the classification tags #ABCD, #UYTR, and #NBVC can be set as a decision tree.

For example, as shown in FIG. 7, the data vector associated with the classification tag #ABCD may be Data # 1 <0.4, Data # 2 <30, and Data # 3> 150. In addition, there is a relationship of Data # 1, Data # 2, and Data # 3 associated with the classification tag #ABCD in FIG. 7, but these are omitted.

Also, the data vector associated with the classification tag #UYTR may be Data # 1> 0.8, Data # 3> 100, and Data # 3 = 180. In addition, there is a relationship of Data # 1, Data # 2, and Data # 3 associated with the classification tag #UYTR in FIG. 7, but these are omitted.

Also, the data vector associated with the classification tag #NBVC may be Data # 1 = 1.2, Data # 1 <10, and Data # 2> 50. In addition, there is a relation of Data # 1, Data # 2, and Data # 3 associated with the classification tag #NBVC in FIG. 7, but these are omitted.

On the other hand, the machine learning model of FIG. 8 uses CLUSTERING in vector space. That is, the data vectors associated with one classification tag may be clustered more closely in the vector space than the data vectors associated with the other classification tag, and thus may be clustered into a single group.

Also, the machine learning model of FIG. 9 can be formed through the relationship between the components of the sequentially generated data vectors included in one classification tag. The data vectors are formed sequentially, and the machine learning can be done through the state change between the components of two consecutive data vectors.

For example, as shown in FIG. 9, the data vectors included in the classification tag #ABCD are (D11, D21, D31), D12, D22, D32, D13, D23, D33, , D34), (D15, D25, D35), (D16, D26, D36).

The state change between D11 and D12, the state change of D12 and D13, the state change of D13 and D14, the state change of D14 and D15, and the state change of D15 and D16.

This state change calculation can be made between D21 and D22, D22 and D23, D23 and D24, D24 and D25, D25 and D26, and likewise D31 and D32, D32 and D33, D33 and D34, D34 and D35, D36. &Lt; / RTI &gt;

The criterion of the state change can be variously set according to the case. For example, if the difference between two consecutive components is greater than 20, the state can be set to change from State # 1 to State # 2. If the ratio between the two consecutive components is greater than 1, it can be set to change from State # 2 to State # 3. A criterion for state change from State # 2 to State # 1, and a state change state from State # 3 to State # 1 can be set.

The ratio of the number of state changes to the number of total state changes can be calculated according to the criterion of the state change, and the ratio can be a model for running the model.

The processor 104 applies the first analysis target time series data and the second analysis target time series data to a machine learning model according to the data vector to classify the first analysis subject time series data and the second analysis subject time series data You can sort by tag.

In other words, as shown in FIGS. 7 to 9, various modeling models can be generated according to the classification tag. In the data searching apparatus according to the embodiment of the present invention, the first analysis target time series data and the second analysis target data Time series data can be input.

The processor 104 may classify the first analysis target time series data and the second analysis target time series data by classification tags by applying the first input analysis target time series data and the second analysis target time series data to the machine learning model.

That is, the processor 104 can apply the first analysis target time series data and the second analysis target time series data to the machine learning model without deriving the matching data and without allocating the comment, 2 Time series data to be analyzed can be classified into specific classification tags.

Next, a data searching apparatus according to another embodiment of the present invention will be described with reference to the drawings.

The data search apparatus according to another embodiment of the present invention includes a memory 106 for storing time series data and a processor 104 that is accessible to the memory 106. [

The processor 104 may assign externally input comments at some time or part of time-series data, and classify the comments according to the classification tags included in the comments.

10 to 12 show comments assigned to the section selected by the user.

As shown in FIG. 10, the user can input a comment into the section of the time series data selected by the user through the input unit 118 or the terminal. Accordingly, the processor 104 can assign the comment entered in the section selected by the user.

At this time, the comment may include a classification tag, and the processor 104 may generate a classification tag list. The classification tag list has been described in detail above, and a description thereof will be omitted.

In Fig. 10, a comment is assigned to the selected section, but may be assigned at the time when the comment is selected, as shown in Fig. In addition, comments are classified according to the classification tag included in the comment, and a classification tag list can be generated.

12, a comment may be assigned to at least one of the selected section or the selected time, the comment may be classified according to the classification tag assigned to the comment, and the classification tag list may be generated.

That is, as shown in FIGS. 10-12, the processor 104 may generate a list of comments on comments and associate them with a classification tag, and generate a classification tag list for the classification tag.

The user can drag a mouse, a stylus, or a touch screen of the terminal to select a specific section or a specific point of time of the specific time series data, and then create and save a comment for the section.

The processor 104 may receive and assign scores for at least one of the classification tag and the comment from one or more user terminals, and may calculate the number of citations of a comment when the comment is cited in another comment. Accordingly, the processor 104 can calculate the price for the comment according to the number of citations and the number of citations to the comment.

Since this has been described above with reference to the data searching apparatus according to the embodiment of the present invention, a description thereof will be omitted.

On the other hand, the processor 104 may generate a data vector having characteristics of time series data assigned a comment, and apply another time series data to the model learning model according to the data vector to classify another time series data according to the classification tag .

The description has been described in detail with reference to the data searching apparatus according to the embodiment of the present invention, and a description thereof will be omitted.

On the other hand, the processor 104 may apply another time series data to the machine learning model without comment assignment. The description has been described in detail with reference to the data searching apparatus according to the embodiment of the present invention, and a description thereof will be omitted.

The machine learning model described above may also be evaluated by the user so that the score for the machine learning model can be assigned by the processor 104. Depending on the score for the machine learning model, The price can be calculated.

The processor 104 can control the process of buying and selling price-estimated machine learning models, and when a machine learning model is sold, the process of compensating the user who built the machine learning model can also be controlled.

It will be apparent to those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit or scope of the invention as defined in the appended claims. . Therefore, the above-described embodiments are to be considered as illustrative rather than restrictive, and the present invention is not limited to the above description, but may be modified within the scope of the appended claims and equivalents thereof.

The processor 104,
The memory 106,
The display unit 114,
The input unit 118,

Claims (5)

A memory for storing time series data; And
A processor accessible to the memory,
The processor comprising:
A comment input by the user is assigned to a part of the time series data or a part of the time series data,
Classifies the comments according to a classification tag included in the comment,
The comment is an interpretation, a comment or a memo of the user with respect to a partial section or a partial view of the time series data,
The processor comprising:
Generates a data vector having characteristics of the time series data to which the comment is assigned,
And classifies the time series data according to the classification tag by applying another time series data to the machine learning model according to the data vector. The method according to claim 1,
The processor comprising:
Generating a comment list for the comment, associating the comment list with the classification tag,
And generates a classification tag list for the classification tag. 3. The method of claim 2,
The processor comprising:
A score for at least one of the classification tag and the comment is received from one or more user terminals,
Calculates the number of citations of the comment when the comment is quoted in another comment,
And calculates a price for the comment according to the score and the number of citations of the comment. delete The method according to claim 1,
The processor comprising:
And applying said another time series data to said machine learning model without comment allocation.

Images