JP2013164811A - Data retrieval device, data retrieval method, and data retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a data retrieval device, a data retrieval method, and a data retrieval program which are capable of reducing a distance calculation cost required for retrieving a pair of pieces of similar partial time-series data from time-series data of a plurality of series that has flowed in within a predetermined time from the past to the present.SOLUTION: An algorithm is employed which has a data structure easy for insertion of partial time-series data at a new receipt time and deletion of partial time-series data at an old receipt time, and discriminates the partial time-series data by the receipt time to prunes them.

Description

  The present invention relates to a data search apparatus, a data search method, and a data search program for searching for a similar pair of partial time-series data from a plurality of series of time-series data that have flowed within a certain period in the past from the present.

  The time series data is a data series observed continuously in time. Time series data is used for environmental observation, equipment monitoring, stock prices, statistical information, media processing, and so on.

  There are two types of time-series data search: searching for data stored in advance and searching for data that flows in sequentially. As an example of the former, there is a process of searching for a similar pattern using precipitation data that has been regularly observed in the past. As an example of the latter, there is a process of observing a stock price change in real time and notifying the moment when it matches a target pattern. The present invention deals with the latter real-time processing.

  The prior art of real-time time series data search includes the following. There has been a search process for obtaining a pair of similar time-series data among a plurality of series of time-series data at the same time (see, for example, Patent Document 1 and Non-Patent Document 1). Patent Document 1 deals with a problem of searching for a pair having a higher degree of similarity than the threshold, and Non-Patent Document 1 searching for a correlation coefficient that is greater than or equal to the threshold. In these prior arts, the calculation cost for similarity determination increases as the number of series increases, and therefore, the technical problem is a dimensional compression for reducing similarity calculation and an index technique for reducing the number of similarity calculations. However, these prior arts have a limitation that only the end point of the time series data is the current time. That is, it cannot be applied to a search for similar partial time-series data having a time lag.

JP 2006-244389 A

Abdullah Muen, Sumna Nat and Jie Liu, “Fast Approximate Correlation for Massive Time-series of Data”, In Proceedings of the 20th Ace. Yianilos, Peter N, “Data structures and algorithms for nearest neighbors, search in general metric spaces, In Proceedings of theSimum in the AUM. Society for Industrial and Applied Materials Philadelphia, pp. 311-321, 1993.

  As described above, the conventional technology cannot handle a search process for obtaining a similar pair of partial time-series data from a plurality of series of time-series data that has flowed in a certain period in the past from the present, that is, time It is not possible to search similar partial time-series data having a misalignment from a plurality of time-series data. As an example of such a search, there is a process of searching for similar partial time-series data from the time-series data of stock prices of a plurality of issues for one month to find characteristic stock price fluctuations with a time lag. Another example of a search is a process of searching for similar partial time-series data from time-series data of a plurality of sensor values to find a phenomenon in which a time lag such as geothermal heat or air temperature occurs from solar radiation. As described above, there are prior arts for multiple series of time series data and prior arts for similar partial time series data with time lag, but time series data search that satisfies these two simultaneously Processing techniques have not been invented.

  The time-series data search process that satisfies the above two simultaneously deals with a search process for obtaining a pair of partial time-series data that is more similar to the threshold value from a plurality of time-series data that has flowed in from a present to a certain period in the past. It will be. This search process is required to reduce the cost of distance calculation, and a reduction in the number of similarity calculations is a technical problem to be solved.

  That is, as the number of time-series data is larger, the time-series data to be searched is longer, or the partial time-series data is shorter than the time-series data, the number of similarity calculations increases. In order to reduce the number of similarity calculations, a search index that narrows down pairs of time-series data having a high similarity is necessary. Although the search index of the prior art used the time series data with the end point of the current time as the search target, in the time series data search processing that satisfies the above two simultaneously, the partial time series data with different reception times is also the search target. However, there is a problem that the calculation cost is increased.

  The present invention has been made to solve the above-mentioned problems, and its purpose is to create a pair of similar partial time-series data from a plurality of series of time-series data that have flowed within a certain period in the past from the present. An object of the present invention is to provide a data search device, a data search method, and a data search program that can reduce the required distance calculation cost when searching.

  The present invention employs an algorithm that easily inserts partial time-series data at a new reception time and deletes partial time-series data at an old reception time, and distinguishes and prune the partial time-series data for each reception time. It was decided. Here, “pruning” means that data not to be searched is removed in advance.

One data search device according to the present invention includes a receiving unit that receives time-series data of a plurality of sequences that sequentially flows,
A first memory for storing the time-series data received by the receiving unit;
An interval dividing unit that reads time-series data from the first memory and divides the time-series data into partial time-series data of a predetermined time;
A similarity determination unit for obtaining a similarity between the partial time-series data divided by the section dividing unit;
Based on the similarity obtained by the similarity determination unit, an index unit that creates a similarity search index of the partial time-series data generated by the section division unit;
A second memory for storing the similar search index created by the index unit in a queue structure for each reception time;
Using the partial time-series data generated by the section dividing unit as a query, a similar pair search unit that reads partial time-series data similar to a specified threshold value from the second memory using the similar search index;
Is provided.

  One data search method according to the present invention divides time-series data of a plurality of sequences that are sequentially input into partial time-series data of a predetermined time, obtains the similarity between the partial time-series data, and determines the similarity A similar search index is created and stored in a queue structure for each reception time, the partial time series data is used as a query at the time of search, and the partial time series data similar to a specified threshold is used as the query Read out.

  The present invention creates a similar search index for each reception time, thereby speeding up deletion of partial time-series data of reception times outside the specified time range from the index and addition of new partial time-series data to the index, and Therefore, pruning can be performed by distinguishing partial time-series data for each reception time. Therefore, according to the present invention, when the number of series of time series data is large, it is possible to increase the speed of updating partial time series data and reduce the number of times of similarity calculation.

  Therefore, according to the present invention, it is possible to reduce the distance calculation cost required when searching for a pair of similar partial time-series data from a plurality of time-series data that has flowed in from the present within a certain period in the past. A data search device and a data search method can be provided.

Another data search device according to the present invention, a receiving unit that receives time-series data of a plurality of series that sequentially flows,
A first memory for storing the time-series data received by the receiving unit;
An interval dividing unit that reads time-series data from the first memory and divides the time-series data into partial time-series data of a predetermined time;
A similarity determination unit for obtaining a similarity between the partial time-series data divided by the section dividing unit;
An index that has a structure in which the partial time-series data can be distinguished by reception time, and creates a similarity search index for the partial time-series data generated by the section division unit based on the similarity obtained by the similarity determination unit And
A second memory for storing the similar search index created by the index unit and adding the subsequent partial time series data as a queue to the similar search index while maintaining the structure of the similar search index;
Using the partial time-series data generated by the section dividing unit as a query, a similar pair search unit that reads partial time-series data similar to a specified threshold value from the second memory using the similar search index;
Is provided.

  Another data search method according to the present invention divides time-series data of a plurality of series that are sequentially input into partial time-series data of a predetermined time, obtains a similarity between the partial time-series data, and the partial time-series data The data has a structure that can be distinguished by reception time, and creates and stores one similar search index based on the similarity, and the subsequent partial time-series data is maintained while maintaining the structure of the similar search index. A queue is added to the similar search index, and the partial time series data similar to a specified threshold is read using the partial search data as a query at the time of search using the similar search index.

  In the present invention, a plurality of similar search indexes created at reception times within a specified time range are obtained by making the similar search indexes into a structure in which each time-series data can be distinguished by reception time and then consolidating them into one. It is possible to reduce the time and effort of searching for each search. Therefore, the present invention can realize a reduction in the number of times of similarity calculation when the time series data to be searched is long or when the partial time series data is short relative to the length of the time series data. .

  Therefore, according to the present invention, it is possible to reduce the distance calculation cost required when searching for a pair of similar partial time-series data from a plurality of time-series data that has flowed in from the present within a certain period in the past. A data search device and a data search method can be provided.

  A data search program according to the present invention causes a computer to execute the data search method.

  The present invention provides a data search that can reduce the required distance calculation cost when searching for a pair of similar partial time-series data from a plurality of time-series data that has flowed within a certain period in the past from the present. An apparatus and a data search method can be provided.

It is a figure of the image explaining a similar partial time series pair. It is a figure explaining the data search device concerning the present invention. It is a figure explaining the data search method which concerns on this invention. It is a figure of an image when the data search device concerning the present invention divides time series data into partial time series data. It is a figure explaining the queue of the similar search index in the data search method concerning the present invention. It is a figure explaining the similar search index structure in the data search method which concerns on this invention. It is a figure of the image of pruning using the similar search index in the data search method which concerns on this invention. It is a figure explaining the data search device concerning the present invention. It is a figure explaining the data search method which concerns on this invention. It is a figure explaining the queue of the similar search index in the data search method concerning the present invention.

  Embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments described below are examples of the present invention, and the present invention is not limited to the following embodiments. In the present specification and drawings, the same reference numerals denote the same components.

  FIG. 1 is a diagram of an image of similar partial time series pairs. In the apparatus according to the first and second embodiments, a partial time series data pair that is more similar than the threshold is searched.

(Embodiment 1)
FIG. 2 is a diagram illustrating the data search apparatus 301 according to the present embodiment. The data search device 301 includes a receiving unit 11 that receives time-series data of a plurality of sequences that sequentially flows,
A first memory 12 for storing the time-series data received by the receiving unit 11;
An interval division unit 13 that reads time-series data from the first memory 12 and divides the time-series data into partial time-series data of a predetermined time;
A similarity determination unit 14 for obtaining a similarity between the partial time-series data divided by the section dividing unit 13;
An index unit 15 that creates a similarity search index of the partial time-series data generated by the section division unit 13 based on the similarity obtained by the similarity determination unit 14;
A second memory 16 for storing the similar search index created by the index unit 15 in a queue structure for each reception time;
A similar pair search unit 17 that reads out the partial time-series data similar to a specified threshold value from the second memory 16 using the similar search index, using the partial time-series data generated by the section dividing unit 13 as a query;
Is provided.

  FIG. 3 is a diagram for explaining a data search method performed by the data search apparatus 301. The data search method divides time series data of a plurality of series that are sequentially input into partial time series data of a predetermined time (steps S101 and S102), obtains the similarity between the partial time series data, and calculates the similarity A similar search index based on the above is created and stored in a queue structure for each reception time (step S103), and the partial time series data similar to a specified threshold is used as a query at the time of search. Reading is performed using the similar search index (step S104).

  In step S <b> 101, the receiving unit 11 receives a plurality of time series data that sequentially flows and stores the time series data in the first memory 12. When the capacity of the first memory 12 is limited, the time series data of the reception time outside the designated time range is deleted from the first memory 12. The minimum required storage capacity of the first memory 12 is the number of elements of n (m + w−1), where m is the partial time series data length to be searched, w is the time length to be searched, and n is the number of series. It becomes the capacity which can be stored.

は受信時刻ｔ_ｘにおける系列のＩＤがｉの部分時系列データを表している。 In step S102, the section dividing unit 13 reads the time series data from the first memory 12, and divides it into partial time series data of the designated section time. The partial time series data in the present embodiment refers to a continuous data series in a specified section time among temporally continuous data series that sequentially flow in. Every time new time series data is received, the section division unit 13 divides the partial time series data of the most recent length m of each time series data, and the partial time series data is sent to the index unit 15 and the similar pair search unit 17. input. In addition, the partial time-series data of the length m of time specified from the similar pair search unit 17 is divided and returned. FIG. 4 is a diagram of an image when the section dividing unit 13 divides time series data into partial time series data. In the figure

The ID of the sequences in reception time t _x represents a partial time-series data of i.

の類似度Ｄｉｓｔにユークリッド距離を採用し、以下の通り定義する。

In step S103, the similarity determination unit 14 calculates the similarity between the partial time series data. The similarity in the present embodiment is an index that expresses how similar two partial time-series data are numerically. In particular, in this embodiment, two partial time series data

The Euclidean distance is adopted as the similarity degree Dist and is defined as follows.

  It should be noted that the similarity referred to in the present invention is subject to various distances that satisfy the above characteristics, and is not limited to the Euclidean distance. Specifically, for example, distances representing similarity between time series data including Dynamic time warping, Earth Move's distance, editing distance, Hamming distance, Jaccard coefficient distance, cosine similarity, Manhattan distance, Mahalanobis distance, etc. It may be. Furthermore, in addition to these distances, distances to those adjusted for partial time series offset and amplitude may be used.

  In step S103, the index unit 15 manages a plurality of similar search indexes generated at the reception time within the time range specified in the second memory 16 in a queue structure. The similarity search index in the present embodiment is an accumulated partial time series in which the similarity with the query is within a threshold without determining all the similarities between the partial time series data given as a query and the accumulated partial time series data. A data structure for searching for data. The index unit 15 newly creates a similar search index based on the partial time-series data of the latest reception time divided by the section dividing unit 13 and adds the similar search index to the queue in which the index of the second memory 16 is accumulated. At the same time, the similar search index of the reception time outside the designated time range is deleted from the queue storing the index of the second memory 16. FIG. 5 is a diagram for explaining a queue of similar search indexes.

の２つの部分時系列データ集合Ｓ_ｉｎとＳ_ｏｕｔに分類される。Ｓ_ｉｎとＳ_ｏｕｔも同様にそれぞれで基準となる部分時系列データと分割距離を決めてさらに空間を２分類することを繰り返すことで、木構造の類似検索索引をつくる。この類似検索索引は、葉ノードと内部ノードという２種類のノードで構成される。葉ノードは木構造の末端のノードであり、内部ノードは末端以外に位置するノードをいう。分類のたびに必要となるＳ^ｐ _ｔとｒ^ｐ _ｔは、内部ノードを生成して管理する。十分に小さい数の部分時系列データ集合まで繰り返し分類して必要な数の内部ノードを生成した後は、葉ノードを生成してＳ_ｔを管理する。つまり、内部ノードは分類に必要な情報（Ｓ^ｐ _ｔ、ｒ^ｐ _ｔ）を管理し、葉ノードでは部分時系列データ集合を管理する。図６は、本実施形態の類似検索索引構造を説明する図である。 [How to create a similar search index]
The similarity search index in this embodiment employs a similarity search index Vantage Point Tree (VPT) (see Non-Patent Document 2) in the metric space, and is created by the following procedure. This index defines one partial time series data (reference partial time series data) and one division distance as a reference, and a plurality of partial time series to be indexed based on the distance and division distance from the reference partial time series data Create data by repeatedly classifying data into two. Here, an example of 2 classify partial time-series data set S _t of reception time t. The partial time-series data ^S _{p t} in relation to the standard sequence p select one from _{S t,} and the distance that can be evenly divided into two _{S t} was ^r _{p t, S t} is

_Are classified into two partial time series data sets S _in and S _out . Similarly for S _in and S _out , a similar search index having a tree structure is created by repeatedly determining the partial time series data and the division distance as reference and further classifying the space into two. This similar search index is composed of two types of nodes, leaf nodes and internal nodes. The leaf node is a node at the end of the tree structure, and the internal node is a node located at a position other than the end. S ^p _t and r ^p _t required for each classification are generated by managing internal nodes. After iteratively classifying a sufficiently small number of partial time-series data sets and generating the required number of internal nodes, leaf nodes are generated and _St is managed. That is, the internal node manages information (S ^p _t , r ^p _t ) necessary for classification, and the leaf node manages a partial time-series data set. FIG. 6 is a diagram for explaining the similar search index structure of the present embodiment.

  Note that the similar search index referred to in the present invention is intended for various data structures that satisfy the above characteristics, and is not limited to the VPT. Specific examples include similar search indexes for other metric spaces such as M-tree, PM-tree, MVP-tree, sa-tree, iDistance, List of Clusters, AESA, LAESA, GNAT, D-index, etc. . Furthermore, in addition to these metric space similar search indexes, there are similar search indexes R-tree and Priority R-tree for multi-dimensional spaces, similar search indexes for character string data, and similar search indexes for set data. May be. Further, the scope of the present invention is not limited to this, and it is possible to combine a dimensional compression technique with these indexes. Specific examples include DFT, SVD, PAA, SAX and the like. If partial time-series data is long and the pruning is insufficient due to an increase in the amount of data due to an increase in the number of dimensions ("curse of dimension"), the upper and lower limits of distance can be achieved by adopting these dimension compression techniques. Values can be used for indexing.

[Similar data search method]
In step S104, the similar pair search unit 17 uses the latest reception time partial time-series data divided by the section division unit 13 as a query, and the similar search index for each reception time managed in the second memory 16 in a queue structure. Is used to search for partial time series data that is more similar to the query than the threshold. And the similar pair search part 17 outputs a similar partial time series data pair. The data pair here means a pair of partial time-series data at the latest time and similar partial time-series data. The similar pair search unit 17 performs the following procedure and outputs a similar partial time series data pair.

からｒ^ｑの距離の範囲にある時系列データを探したい場合を想定する。数５のクエリと空間分割の基準とした部分時系列データ

との間の距離

が、
１）

のとき、Ｓ_ｉｎに含まれる部分時系列データを検索すれば十分で、Ｓ_ｏｕｔに含まれる時系列データ、つまりＳ_ｏｕｔを管理する内部ノードもしくは葉ノードは枝刈りできる。
２）

のとき、Ｓ_ｏｕｔに含まれる時系列データを検索すれば十分で、Ｓ_ｉｎに含まれる時系列データ、つまりＳ_ｉｎを管理する内部ノードもしくは葉ノードは枝刈りできる。
３） 上記１），２）のどちらの不等式も満たさない場合、Ｓ_ｉｎとＳ_ｏｕｔの両方の時系列データを検索する必要がある。 Here, a search example using a structure in which the above-described S _t is classified into two partial time-series data sets of S _in and S _out by S ^p _t and r ^p _t will be described. Time series data used as a query

It is assumed that you want to find the time-series data is in the range of distances r ^q from. Partial time-series data based on 5 queries and space division

Distance between

But,
1)

In this case, it is sufficient to search the partial time series data included _in S _in, and the time series data included in S _out , that is, the internal node or leaf node that manages S _out can be pruned.
2)

When, it suffices to find the time-series data contained in the S _out, the time-series data contained in the S _in, ie internal nodes or leaf nodes to manage S _in can be pruned.
3) When neither the inequality of 1) nor 2) is satisfied, it is necessary to search both time series data of S _in and S _out .

Such pruning is performed recursively. In other words, the internal node shown in FIG. 6 is traced to some extent, and a search is performed for internal nodes and leaf nodes that cannot be pruned. FIG. 7 is a diagram of an image of pruning using a similar search index. And it satisfies the condition of the above 1), and represents a state capable of pruning time series data contained in the S _out.

  With the above procedure, the data search device 301 searches for a partial time-series data pair that is more similar to the threshold value from a plurality of series of time-series data that has flowed in from the present within a certain period in the past.

(Embodiment 2)
FIG. 8 is a diagram illustrating the data search apparatus 302 according to this embodiment. The data search device 302 includes a receiving unit 11 that receives time-series data of a plurality of sequences that sequentially flows,
A first memory 12 for storing the time-series data received by the receiving unit 11;
An interval division unit 13 that reads time-series data from the first memory 12 and divides the time-series data into partial time-series data of a predetermined time;
A similarity determination unit 14 for obtaining a similarity between the partial time-series data divided by the section dividing unit 13;
An index that has a structure in which the partial time-series data can be distinguished by reception time, and creates a similarity search index for the partial time-series data generated by the section division unit 13 based on the similarity obtained by the similarity determination unit 14 Part 15;
A second memory 16 for storing the similar search index created by the index unit 15 and adding the subsequent partial time series data as a queue to the similar search index while maintaining the structure of the similar search index;
A similar pair search unit 17 that reads out the partial time-series data similar to a specified threshold value from the second memory 16 using the similar search index, using the partial time-series data generated by the section dividing unit 13 as a query;
Is provided.

FIG. 9 is a diagram for explaining a data search method performed by the data search apparatus 302. This data search method divides time-series data of a plurality of series that are sequentially input into partial time-series data of a predetermined time, obtains a similarity between the partial time-series data, and determines the partial time-series data based on a reception time. One similar search index having a distinguishable structure and based on the similarity is created and stored (steps S201 and S202),
The subsequent partial time series data is added as a queue to the similar search index while maintaining the structure of the similar search index (steps S203 to S205),
Using the partial time-series data as a query at the time of search, partial time-series data similar to a specified threshold is read using the similar search index (step S206).

  In the data search device 302, the reception unit 11, the section division unit 13, the similarity determination unit 14, the similar pair search unit 17, and the first memory 12 perform the same operations as the data search device 301 in FIG.

  The index unit 15 builds a similar search index once in step S202, and thereafter updates only the leaf node without changing the internal node (steps S203 to S205). Similar to the first embodiment, the similar search index in the present embodiment adopts the data structure of the similar search index VPT in the metric space.

を“Ｓ”から一つ選び、“Ｓ”を均等２分割できる距離を

とおく。“Ｓ”は、

の２つの部分時系列データ集合Ｓ_ｉｎとＳ_ｏｕｔに分類される。 In steps (S201 to S202), after the number of elements of n (m + w-1) is accumulated in the receiving unit 11, one similar search index is constructed. A partial time series data set that can be cut out by the section dividing unit 13 is assumed to be “S”. Partial time series data with reference series ID p and reception time t _x

Select one from “S” and set the distance that can divide “S” into two equal parts.

far. “S” means

_Are classified into two partial time series data sets S _in and S _out .

S _in the S _out also by repeating two more classifying that the space determined the partial time-series data and dividing the distance as a reference in each similarly, creating an index of the tree structure. After classifying a sufficiently small number of partial time-series data sets, “S” is managed as a leaf node. At this time, the leaf node manages the partial time series data in a queue structure that is easy to access for each reception time.

を管理し、葉ノードでは部分時系列データ集合を管理する。図１０は、類似検索索引の葉ノードにおけるキューを説明する図である。 In this way, in the internal node of the tree structure

The leaf node manages a partial time series data set. FIG. 10 is a diagram for explaining a queue in the leaf node of the similar search index.

  The similar search index referred to in the present invention is intended for various data structures satisfying the above characteristics, and is not limited to the VPT, and may be the similar search index shown in the first embodiment. Furthermore, it is possible to combine the dimensional compression technique shown in Embodiment 1 with these indexes.

ならばＳ_ｉｎを担当するノードを、

ならばＳ_ｏｕｔを担当するノードを探索することを繰り返して、該当する葉ノードを探索する。そして、葉ノードが見つかったら、部分時系列データを蓄積するキューにＳ^ｉ _ｔを追加する。なお、所定時間が経過した部分時系列データは葉ノードにおけるキューから削除される。 After the index built, the data search apparatus 302 in step S203～S205, the partial time series ^S _{i t} the newly added, to add to the appropriate leaf node. In other words, if there is an internal node that manages the number 13,

If the node in charge of _{S in,}

Then, the search for the node in charge of _Sout is repeated to search for the corresponding leaf node. And, When you find leaf node, add the S ⁱ _t in the queue for accumulating the partial time-series data. Note that the partial time series data for which a predetermined time has elapsed is deleted from the queue in the leaf node.

  Step S206 is the same as step S104 described in FIG. The similar pair search unit 17 uses the partial time-series data of the latest reception time divided by the section division unit 13 as a query and uses the similar search index managed in the second memory 16 to trace the internal nodes in FIG. 6 to some extent. Thus, in the internal node or leaf node that cannot be pruned, partial time series data that is more similar to the query than the threshold is searched. And the similar pair search part 17 outputs a similar partial time series data pair.

  With the above procedure, the data search device 302 searches for partial time-series data pairs that are more similar to the threshold value from a plurality of series of time-series data that have flowed in from the present within a certain period in the past.

11: receiving unit 12: first memory 13: section dividing unit 14: similarity determining unit 15: index unit 16: second memory 17: similar pair determining unit 18: third memories 301 and 302: data search device

Claims (5)

A receiving unit for receiving time series data of a plurality of series that sequentially flows;
A first memory for storing the time-series data received by the receiving unit;
An interval dividing unit that reads time-series data from the first memory and divides the time-series data into partial time-series data of a predetermined time;
A similarity determination unit for obtaining a similarity between the partial time-series data divided by the section dividing unit;
Based on the similarity obtained by the similarity determination unit, an index unit that creates a similarity search index of the partial time-series data generated by the section division unit;
A second memory for storing the similar search index created by the index unit in a queue structure for each reception time;
Using the partial time-series data generated by the section dividing unit as a query, a similar pair search unit that reads partial time-series data similar to a specified threshold value from the second memory using the similar search index;
A data search device comprising: A receiving unit for receiving time series data of a plurality of series that sequentially flows;
A first memory for storing the time-series data received by the receiving unit;
An interval dividing unit that reads time-series data from the first memory and divides the time-series data into partial time-series data of a predetermined time;
A similarity determination unit for obtaining a similarity between the partial time-series data divided by the section dividing unit;
An index that has a structure in which the partial time-series data can be distinguished by reception time, and creates a similarity search index for the partial time-series data generated by the section division unit based on the similarity obtained by the similarity determination unit And
A second memory for storing the similar search index created by the index unit and adding the subsequent partial time series data as a queue to the similar search index while maintaining the structure of the similar search index;
Using the partial time-series data generated by the section dividing unit as a query, a similar pair search unit that reads partial time-series data similar to a specified threshold value from the second memory using the similar search index;
A data search device comprising: A plurality of time series data that are sequentially input are divided into partial time series data of a predetermined time, the similarities of the partial time series data are obtained, and a similar search index based on the similarity is created for each reception time. In the queue structure,
A data search method for reading out partial time-series data similar to a specified threshold using the similar search index, using the partial time-series data as a query at the time of search. The time series data of a plurality of series that are sequentially input is divided into partial time series data of a predetermined time, a similarity between the partial time series data is obtained, and the partial time series data can be distinguished by reception time. One similarity search index based on the similarity is created and stored,
The subsequent partial time series data is added as a queue to the similar search index while maintaining the structure of the similar search index,
A data search method for reading out partial time-series data similar to a specified threshold using the similar search index, using the partial time-series data as a query at the time of search.   A data search program for causing a computer to execute the data search method according to claim 3 or 4.

Images