JP2020071751A - Trading examination device, trading examination method, and trading examination program - Google Patents
Trading examination device, trading examination method, and trading examination program Download PDFInfo
- Publication number
- JP2020071751A JP2020071751A JP2018206232A JP2018206232A JP2020071751A JP 2020071751 A JP2020071751 A JP 2020071751A JP 2018206232 A JP2018206232 A JP 2018206232A JP 2018206232 A JP2018206232 A JP 2018206232A JP 2020071751 A JP2020071751 A JP 2020071751A
- Authority
- JP
- Japan
- Prior art keywords
- data
- transaction data
- suspected
- transaction
- trading
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
Description
本発明は、売買審査装置、売買審査方法及び売買審査プログラムに関する。 The present invention relates to a trade examination device, a trade examination method, and a trade examination program.
従来、株式市場等の金融商品市場における不公正取引の疑いがある取引を検出するため、売買審査が行われている。売買審査は、不公正取引が疑われる事案を抽出する調査段階と、抽出された事案を詳細に分析する審査段階を含む。 Conventionally, trading screening is performed to detect transactions that are suspected of being unfair in the financial instruments markets such as the stock market. The trading examination includes an examination stage for extracting cases suspected of being unfair trade and an examination stage for analyzing the extracted cases in detail.
調査段階を自動化する試みとして、下記非特許文献１には、不公正取引の疑いがある取引に関するデータであるか否かを示すラベルを取引データに付与して、不公正取引が疑われる事案を抽出するモデルを教師有り学習によって生成する技術が記載されている。 As an attempt to automate the investigation stage, in Non-Patent Document 1 below, a case is shown in which a label indicating whether or not the data is related to a transaction suspected of being unfair transaction is given to the transaction data, and the unfair transaction is suspected. A technique for generating a model to be extracted by supervised learning is described.
非特許文献１に記載の技術では、不公正取引の疑いがある取引に関するデータであるか否かを示すラベルを取引データに付与して教師データを作成する必要がある。そのため、教師データを作成するコストが過大となる。 In the technique described in Non-Patent Document 1, it is necessary to create a teacher data by adding a label indicating whether or not the data is related to a transaction suspected of being unfair transaction to the transaction data. Therefore, the cost of creating the teacher data becomes excessive.
また、予め定められた抽出基準に従って不公正取引が疑われる取引を抽出する方法が実務で採用されているが、これは、市場全体を統一的に取り扱っているものの、売買高２０％関与率等の数値的な基準に基づく抽出であり、不公正取引の疑いのある取引も疑いのない取引も同時に抽出され、抽出結果は不公正取引の疑いのない取引が大部分を占めている。従って、既存の手法では、不公正取引の疑いのある取引を効率良く抽出することが困難である。 In practice, a method of extracting transactions that are suspected of being unfair according to a predetermined extraction standard is adopted in practice, but this involves handling the entire market in a unified manner, but with a trading volume of 20% involvement rate, etc. This is an extraction based on the numerical standard of, and both transactions with and without suspicion of unfair transactions are extracted at the same time, and the extraction results are mostly transactions without suspicion of unfair transactions. Therefore, it is difficult for existing methods to efficiently extract transactions that are suspected of being unfair.
そこで、本発明は、教師データを必要としない学習モデルによって不公正取引が疑われる取引データであるか判定する売買審査装置、売買審査方法及び売買審査プログラムを提供する。 Therefore, the present invention provides a trading examination device, a trading examination method, and a trading examination program for judging whether the transaction data is a transaction data suspected of being unfair by a learning model that does not require teacher data.
本発明の一態様に係る売買審査装置は、金融商品の売買に関する第１取引データと、金融商品の売買に関し、第１取引データと異なる第２取引データとに基づいて、密度比関数により第１取引データの確率分布と第２取引データの確率分布との密度比を推定する推定部と、密度比関数に基づき、第２取引データの異常度を算出する算出部と、異常度に基づき、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する第１判定部と、を備える。 A trading examination apparatus according to one aspect of the present invention uses a density ratio function to perform a first transaction data based on first transaction data relating to a financial instrument transaction and second transaction data relating to a financial instrument transaction. An estimation unit that estimates the density ratio between the probability distribution of the transaction data and the probability distribution of the second transaction data, a calculation unit that calculates the abnormality degree of the second transaction data based on the density ratio function, and a calculation unit that calculates the abnormality degree based on the abnormality degree. 2 a first determination unit for determining whether or not the transaction data includes first suspected data suspected of being unfair.
この態様によれば、第１取引データの確率分布と第２取引データの確率分布との密度比を推定することで、第１取引データの確率分布と第２取引データの確率分布とを直接推定する必要がなくなり、教師データを必要としない学習モデルによって第２取引データの異常度を算出することができる。これにより、教師データを必要としない学習モデルによって、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することができる。 According to this aspect, by estimating the density ratio between the probability distribution of the first transaction data and the probability distribution of the second transaction data, the probability distribution of the first transaction data and the probability distribution of the second transaction data are directly estimated. The learning model that does not require the teacher data can be used to calculate the abnormality degree of the second transaction data. Accordingly, it is possible to determine whether or not the second transaction data includes the first suspected data suspected of being unfair transaction by the learning model that does not require the teacher data.
上記態様において、第１取引データは、第２取引データよりも長い期間の取引データであってもよい。 In the above aspect, the first transaction data may be transaction data for a longer period than the second transaction data.
この態様によれば、第１取引データの期間が第２取引データの期間よりも長いことで、第１取引データの確率分布によって通常の取引の確率分布を表すことができ、第２取引データの異常度を適切に算出することができる。 According to this aspect, the period of the first transaction data is longer than the period of the second transaction data, and thus the probability distribution of the normal transaction can be represented by the probability distribution of the first transaction data. The degree of abnormality can be calculated appropriately.
上記態様において、第１取引データ及び第２取引データは、所定期間の売買注文量に関するデータを含んでもよい。 In the above aspect, the first transaction data and the second transaction data may include data relating to the sales order quantity for a predetermined period.
この態様によれば、売買注文量の急増や急減を捉えることができ、第２取引データに見せ玉等の不公正取引が疑われる第１被疑データが含まれるか判定することができる。 According to this aspect, it is possible to catch a sharp increase or a sharp decrease in the sales order quantity, and it is possible to determine whether or not the second transaction data includes the first suspected data that is suspected of unfair transaction such as a show ball.
上記態様において、第１取引データ及び第２取引データは、所定期間の売買注文のキャンセル量に関するデータを含んでもよい。 In the above aspect, the first transaction data and the second transaction data may include data relating to the cancellation amount of the buying and selling order for a predetermined period.
この態様によれば、売買注文のキャンセル量の急増や急減を捉えることができ、第２取引データに見せ玉等の不公正取引が疑われる第１被疑データが含まれるか判定することができる。 According to this aspect, it is possible to catch a sudden increase or decrease in the amount of cancellation of the sales order, and it is possible to determine whether or not the second transaction data includes the first suspected data suspected of unfair transaction such as a show ball.
上記態様において、第１判定部は、第２取引データに含まれるデータのうち、異常度が上位所定割合に属するデータを第１被疑データと判定してもよい。 In the above aspect, the first determination unit may determine, as the first suspected data, data having an abnormality degree belonging to a higher predetermined ratio among the data included in the second transaction data.
この態様によれば、異常度について絶対的な閾値を設定することなく、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することができる。 According to this aspect, it is possible to determine whether or not the second transaction data includes the first suspected data suspected of being unfair transaction, without setting an absolute threshold for the abnormality degree.
上記態様において、第１判定部は、第２取引データに含まれるデータのうち、異常度が上位所定割合に属し、所定期間の値幅が第１閾値以上であるデータを第１被疑データと判定してもよい。 In the above aspect, the first determination unit determines, as the first suspected data, data in which the degree of abnormality belongs to the upper predetermined ratio of the data included in the second transaction data and the price range of the predetermined period is equal to or more than the first threshold value. May be.
この態様によれば、異常度について絶対的な閾値を設定せずに、値幅について第１閾値を設定することで、第２取引データに不公正取引が疑われる第１被疑データが含まれるかより高精度に判定することができる。 According to this aspect, by setting the first threshold value for the price range without setting the absolute threshold value for the abnormality degree, it is possible to determine whether the second transaction data includes the first suspected data suspected of being unfair. It can be determined with high accuracy.
上記態様において、第１判定部は、第２取引データに含まれるデータのうち、異常度が上位所定割合に属し、所定期間の出来高が第２閾値以上であるデータを第１被疑データと判定してもよい。 In the above aspect, the first determination unit determines, as the first suspected data, data in which the degree of abnormality belongs to the upper predetermined ratio of the data included in the second transaction data and the output of the predetermined period is the second threshold value or more. May be.
この態様によれば、異常度について絶対的な閾値を設定せずに、出来高について第２閾値を設定することで、第２取引データに不公正取引が疑われる第１被疑データが含まれるかより高精度に判定することができる。 According to this aspect, by setting the second threshold value for the output amount without setting the absolute threshold value for the abnormality degree, it is possible to determine whether the second transaction data includes the first suspected data suspected of unfair transaction. It can be determined with high accuracy.
上記態様において、異常度の時間変化を示す第１グラフを生成するグラフ生成部をさらに備えてもよい。 In the above aspect, a graph generation unit that generates a first graph showing a temporal change in the degree of abnormality may be further included.
この態様によれば、異常度の時間変化が直感的に把握できるようになり、不公正取引が疑われる期間が一目で把握できるようになる。 According to this aspect, it is possible to intuitively grasp the time change of the abnormality degree, and it is possible to grasp at a glance the period in which the unfair transaction is suspected.
上記態様において、グラフ生成部は、第２取引データの時間変化を示し、第２取引データのうち第１被疑データを識別可能に示す第２グラフをさらに生成してもよい。 In the above aspect, the graph generation unit may further generate a second graph showing the time change of the second transaction data and identifiable the first suspected data of the second transaction data.
この態様によれば、第２取引データのうち第１被疑データを識別可能に示す第２グラフによって、不公正取引が疑われる箇所が一目で把握できるようになる。 According to this aspect, the second graph showing the first suspected data in the second transaction data in a distinguishable manner makes it possible to grasp at a glance the location where the unfair transaction is suspected.
上記態様において、予め定められた基準に基づき、第２取引データに不公正取引が疑われる第２被疑データが含まれるか判定する第２判定部をさらに備え、第２グラフは、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示してもよい。 In the above aspect, a second determination unit that determines whether the second transaction data includes second suspected data suspected of being unfair transaction based on a predetermined standard is further provided, and the second graph shows the first suspected data. , The second suspected data and the overlap between the first suspected data and the second suspected data may be shown in a distinguishable manner.
この態様によれば、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、学習モデルによって抽出された第１被疑データと、予め定められた基準に基づきで抽出された第２被疑データとの共通点と相違点を一目で把握することができる。 According to this aspect, the first suspected data, the second suspected data, and the first suspected data extracted by the learning model and the predetermined suspected data can be identified by indicating the overlap between the first suspected data and the second suspected data in a distinguishable manner. It is possible to grasp at a glance the common points and differences with the second suspected data extracted based on the above criteria.
上記態様において、第２取引データの時間変化を示し、第２取引データのうち第１被疑データを識別可能に示す取引明細表を生成する明細表生成部をさらに備えてもよい。 In the above aspect, a schedule table generation unit may be further provided that generates a transaction schedule table that shows a time change of the second transaction data and that can identify the first suspected data in the second transaction data.
この態様によれば、第２取引データのうち第１被疑データを識別可能に示す取引明細表によって、不公正取引が疑われる箇所が一目で把握できるようになる。 According to this aspect, the portion where the unfair transaction is suspected can be grasped at a glance by the transaction details table showing the first suspected data in the second transaction data in a distinguishable manner.
上記態様において、予め定められた基準に基づき、第２取引データに不公正取引が疑われる第２被疑データが含まれるか判定する第２判定部をさらに備え、取引明細表は、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示してもよい。 In the above aspect, a second determination unit for determining whether or not the second transaction data includes second suspected data suspected of being unfair transaction is further provided based on a predetermined standard, and the transaction schedule is the first suspected data. , The second suspected data and the overlap between the first suspected data and the second suspected data may be shown in a distinguishable manner.
この態様によれば、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、学習モデルによって抽出された第１被疑データと、予め定められた基準に基づき抽出された第２被疑データとの共通点と相違点を一目で把握することができる。 According to this aspect, the first suspected data, the second suspected data, and the first suspected data extracted by the learning model and the predetermined suspected data can be identified by indicating the overlap between the first suspected data and the second suspected data in a distinguishable manner. It is possible to grasp at a glance the common points and the differences with the second suspect data extracted based on the above criteria.
本発明の他の態様に係る売買審査方法は、金融商品の売買に関する第１取引データと、金融商品の売買に関し、第１取引データと異なる第２取引データとに基づいて、密度比関数により第１取引データの確率分布と第２取引データの確率分布との密度比を推定することと、密度比関数に基づき、第２取引データの異常度を算出することと、異常度に基づき、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することと、を含む。 A trading examination method according to another aspect of the present invention uses a density ratio function based on first transaction data relating to the trading of financial products and second transaction data relating to the trading of financial products, which is different from the first trading data. 1 estimating the density ratio between the probability distribution of the transaction data and the probability distribution of the second transaction data, calculating the abnormality degree of the second transaction data based on the density ratio function, and calculating the abnormality degree of the second transaction data Determining whether the transaction data includes first suspect data suspected of being unfair.
この態様によれば、第１取引データの確率分布と第２取引データの確率分布との密度比を推定することで、第１取引データの確率分布と第２取引データの確率分布とを直接推定する必要がなくなり、教師データを必要としない学習モデルによって第２取引データの異常度を算出することができる。これにより、教師データを必要としない学習モデルによって、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することができる。 According to this aspect, by estimating the density ratio between the probability distribution of the first transaction data and the probability distribution of the second transaction data, the probability distribution of the first transaction data and the probability distribution of the second transaction data are directly estimated. The learning model that does not require the teacher data can be used to calculate the abnormality degree of the second transaction data. Accordingly, it is possible to determine whether or not the second transaction data includes the first suspected data suspected of being unfair transaction by the learning model that does not require the teacher data.
本発明の他の態様に係る売買審査プログラムは、金融商品の売買に関する第１取引データと、金融商品の売買に関し、第１取引データと異なる第２取引データとに基づいて、密度比関数により第１取引データの確率分布と第２取引データの確率分布との密度比を推定する推定部、密度比関数に基づき、第２取引データの異常度を算出する算出部、及び異常度に基づき、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する第１判定部、として機能させる。 A trading examination program according to another aspect of the present invention uses a density ratio function based on first transaction data relating to the trading of financial products and second transaction data relating to the trading of financial products, which is different from the first trading data. The estimation unit that estimates the density ratio between the probability distribution of the first transaction data and the probability distribution of the second transaction data, the calculation unit that calculates the abnormality degree of the second transaction data based on the density ratio function, and the abnormality unit based on the abnormality degree (2) It functions as a first determination unit that determines whether or not the transaction data includes first suspected data suspected of being unfair.
この態様によれば、第１取引データの確率分布と第２取引データの確率分布との密度比を推定することで、第１取引データの確率分布と第２取引データの確率分布とを直接推定する必要がなくなり、教師データを必要としない学習モデルによって第２取引データの異常度を算出することができる。これにより、教師データを必要としない学習モデルによって、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することができる。 According to this aspect, by estimating the density ratio between the probability distribution of the first transaction data and the probability distribution of the second transaction data, the probability distribution of the first transaction data and the probability distribution of the second transaction data are directly estimated. The learning model that does not require the teacher data can be used to calculate the abnormality degree of the second transaction data. Accordingly, it is possible to determine whether or not the second transaction data includes the first suspected data suspected of being unfair transaction by the learning model that does not require the teacher data.
本発明によれば、教師データを必要としない学習モデルによって不公正取引が疑われる取引データであるか判定する売買審査装置、売買審査方法及び売買審査プログラムが提供される。 According to the present invention, there is provided a trading examination device, a trading examination method, and a trading examination program for judging whether the transaction data is a transaction data suspected of being unfair by a learning model that does not require teacher data.
以下、本発明の一側面に係る実施の形態（以下、「本実施形態」と表記する。）を、図面に基づいて説明する。なお、各図において、同一の符号を付したものは、同一又は同様の構成を有する。 Hereinafter, an embodiment according to one aspect of the present invention (hereinafter referred to as “the present embodiment”) will be described with reference to the drawings. In addition, in each of the drawings, components denoted by the same reference numerals have the same or similar configurations.
図１は、本発明の実施形態に係る売買審査システム１００の概要を示す図である。売買審査システム１００は、売買審査装置１０、第１ユーザ端末２０、第２ユーザ端末３０、証券会社データベース４０及び情報ベンダデータベース５０を備える。なお、同図に示す構成は、売買審査システム１００の一例であり、売買審査システム１００は、これ以外の構成を含んでもよいし、この中の一部の構成を備えなくてもよい。 FIG. 1 is a diagram showing an outline of a trading examination system 100 according to an embodiment of the present invention. The trade examination system 100 includes a trade examination device 10, a first user terminal 20, a second user terminal 30, a securities company database 40, and an information vendor database 50. It should be noted that the configuration shown in the figure is an example of the trading examination system 100, and the trading examination system 100 may include other configurations, or may not include some of the configurations.
売買審査装置１０は、金融商品の売買に関する取引データを解析し、取引データに不公正取引が疑われる被疑データが含まれるか判定する。不公正取引としては、インサイダー取引、相場操縦取引、風説の流布、仮名取引等が含まれる。本実施形態では、売買審査装置１０によって、取引データに見せ玉による相場操縦取引が疑われる被疑データが含まれるか判定する場合について説明する。 The transaction inspection device 10 analyzes transaction data relating to the purchase and sale of financial products, and determines whether the transaction data includes suspected data that is suspected of being unfair. Unfair trading includes insider trading, market manipulation trading, dissemination of rumor, and pseudonym trading. In the present embodiment, a case will be described in which the trading examination device 10 determines whether or not the transaction data includes suspicious data that is suspected of being a market manipulation transaction using a show ball.
第１ユーザ端末２０及び第２ユーザ端末３０は、売買審査を行う事業者が用いる端末である。第１ユーザ端末２０及び第２ユーザ端末３０は、それぞれ独立に、異なる取引データについて不公正取引が行われていないか審査するため、売買審査装置１０に対して被疑データの抽出を依頼する。 The first user terminal 20 and the second user terminal 30 are terminals used by a business operator who conducts trade examination. The first user terminal 20 and the second user terminal 30 each independently request the trading inspecting apparatus 10 to extract suspected data in order to inspect whether or not unfair transactions are performed on different transaction data.
証券会社データベース４０は、顧客データＤＢ１及び取引データＤＢ２を有する。証券会社データベース４０は、顧客に関する情報を顧客データＤＢ１に記憶し、顧客の識別情報と関連付けて、その顧客の取引データを取引データＤＢ２に記憶している。証券会社データベース４０は、必要に応じて匿名加工処理を施した後、取引データを売買審査装置１０に提供する。 The securities company database 40 has a customer data DB1 and a transaction data DB2. The securities company database 40 stores information regarding customers in the customer data DB1, stores the transaction data of the customers in the transaction data DB2 in association with the customer identification information. The securities company database 40 provides transaction data to the trading examination device 10 after anonymizing processing as necessary.
情報ベンダデータベース５０は、マーケットデータＤＢ３を有する。情報ベンダデータベース５０は、金融商品の気配情報及び注文情報（以下「板情報」という。）並びに約定情報等の相場情報を時系列データとしてマーケットデータＤＢ３に記憶している。情報ベンダデータベース５０は、任意の時間間隔で記録された金融商品の板情報及び相場情報を売買審査装置１０に提供する。 The information vendor database 50 has a market data DB3. The information vendor database 50 stores market information DB 3 as market information such as quote information and order information (hereinafter referred to as “board information”) of financial products and contract information as time series data. The information vendor database 50 provides the trading examination device 10 with the board information and the market price information of the financial product recorded at an arbitrary time interval.
図２は、本実施形態に係る売買審査装置１０の機能ブロックを示す図である。売買審査装置１０は、取得部１１、推定部１２、算出部１３、第１判定部１４、第２判定部１５、明細表生成部１６及びグラフ生成部１７を備える。 FIG. 2 is a diagram showing functional blocks of the trading inspection device 10 according to the present embodiment. The trading inspection device 10 includes an acquisition unit 11, an estimation unit 12, a calculation unit 13, a first determination unit 14, a second determination unit 15, a schedule table generation unit 16 and a graph generation unit 17.
取得部１１は、顧客データＤＢ１、取引データＤＢ２及びマーケットデータＤＢ３から、顧客情報、取引データ及びマーケットデータを取得する。なお、取得部１１は、少なくとも金融商品の板情報及び相場情報を含む取引データを取得し、顧客情報については必ずしも取得しなくてもよい。 The acquisition unit 11 acquires customer information, transaction data, and market data from the customer data DB1, transaction data DB2, and market data DB3. The acquisition unit 11 may acquire transaction data including at least board information and market price information of financial products, and may not necessarily acquire customer information.
推定部１２は、金融商品の売買に関する第１取引データと、金融商品の売買に関し、第１取引データと異なる第２取引データとに基づいて、密度比関数により第１取引データの確率分布と第２取引データの確率分布との密度比を推定する。ここで、第１取引データは、第２取引データよりも長い期間の取引データであってよい。例えば、第２取引データは、株式の一日分の取引データであってよく、第１取引データは、それ以前一週間分の同じ株式に関する取引データであってよい。第１取引データ及び第２取引データは、無作為に選択されてよい。その場合、第１取引データに不公正取引が疑われる取引データが含まれる場合があり得るが、第１取引データの期間が第２取引データの期間よりも長いことで、第１取引データの確率分布によって通常の取引の確率分布を近似することができ、第２取引データの異常度を適切に算出することができる。 The estimation unit 12 uses the density ratio function to calculate the probability distribution of the first transaction data based on the first transaction data regarding the transaction of the financial product and the second transaction data regarding the transaction of the financial product, which is different from the first transaction data. 2 Estimate the density ratio with the probability distribution of transaction data. Here, the first transaction data may be transaction data for a longer period than the second transaction data. For example, the second trade data may be one day's worth of stock trade data, and the first trade data may be one week's worth of trade data for the same stock. The first transaction data and the second transaction data may be randomly selected. In that case, the first transaction data may include transaction data suspected of being unfair, but the probability of the first transaction data is that the period of the first transaction data is longer than the period of the second transaction data. The probability distribution of a normal transaction can be approximated by the distribution, and the degree of abnormality of the second transaction data can be calculated appropriately.
第１取引データ及び第２取引データは、所定期間の売買注文量に関するデータを含んでよい。これにより、売買注文量の急増や急減を捉えることができ、第２取引データに見せ玉の不公正取引が疑われる第１被疑データが含まれるか判定することができる。 The first transaction data and the second transaction data may include data relating to the sales order quantity for a predetermined period. Accordingly, it is possible to catch a sharp increase or a sharp decrease in the volume of the sales order, and it is possible to determine whether or not the second transaction data includes the first suspected data suspected of the unfair trade of the show ball.
また、第１取引データ及び第２取引データは、所定期間の売買注文のキャンセル量に関するデータを含んでよい。これにより、売買注文のキャンセル量の急増や急減を捉えることができ、第２取引データに見せ玉等の不公正取引が疑われる第１被疑データが含まれるか判定することができる。 In addition, the first transaction data and the second transaction data may include data regarding the amount of cancellation of the buying and selling order for a predetermined period. This makes it possible to catch a sudden increase or decrease in the amount of cancellation of the sales order, and it is possible to determine whether or not the second transaction data includes the first suspected data suspected of unfair transaction such as a show ball.
第１取引データをｘi（ｉ＝１〜ｎ）と表し、第２取引データをｘ´i（ｉ＝１〜ｎ´）と表し、第１取引データの確率分布をｑ（ｘ）と表し、第２取引データの確率分布をｑ´（ｘ）と表し、密度比関数をｒ（ｘ）＝ｑ´（ｘ）／ｑ（ｘ）と定義する。このとき、推定部１２は、所定の基底関数φk（ｘ）（ｋ＝１〜ｂ）を用いて、ｒ（ｘ）＝Σk=1 bｗkφk（ｘ）のｗkを算出することで、密度比関数を算出してよい。ここで、基底関数は、例えば、第２取引データをｘ´iから無作為にｂ個のデータｃk（ｋ＝１〜ｂ）を抽出し、
φk（ｘ）＝Ｋσ（ｘ，ｃk）と選んでよい。ここで、Ｋσは、標準偏差がσであるガウスカーネルである。
The first transaction data represented as x i (i = 1~n), the second transaction data representing the x'i (i = 1~n'), the probability distribution of the first transaction data represented as q (x) , The probability distribution of the second transaction data is represented as q ′ (x), and the density ratio function is defined as r (x) = q ′ (x) / q (x). At this time, the estimation unit 12 calculates w k of r (x) = Σ k = 1 b w k φ k (x) using a predetermined basis function φ k (x) ( k = 1 to b ). By doing so, the density ratio function may be calculated. Here, the basis function, for example, randomly extracts b pieces of data c k (k = 1 to b) from the second transaction data x ′ i ,
φ k (x) = Kσ (x, c k ) may be selected. Here, Kσ is a Gaussian kernel whose standard deviation is σ.
推定部１２は、密度比関数の二乗誤差を最小化するようにｗkを算出してよい。このとき、ｗkに関する二次の正則化項を追加すると、ｗkの解析解は以下の数式（１）で与えられる。 The estimation unit 12 may calculate w k so as to minimize the squared error of the density ratio function. In this case, adding a secondary regularization term about w k, the analytical solution of w k is given by the following equation (1).
数式（１）において太字で示した量は、ベクトル又は行列であることを表している。ここで、右辺の行列Ｈは、以下の数式（２）で与えられる量である。 The quantities shown in bold in the formula (1) indicate that they are vectors or matrices. Here, the matrix H on the right side is an amount given by the following mathematical expression (2).
すなわち、行列Ｈは、ｂ×ｂ行列であり、基底関数φkに第１取引データｘiを代入した値により構成される。 That is, the matrix H is a b × b matrix, and is composed of values obtained by substituting the first transaction data x i for the basis function φ k .
また、数式（１）の右辺のベクトルｈは、以下の数式（３）で与えられる量である。 Further, the vector h on the right side of Expression (1) is an amount given by Expression (3) below.
すなわち、ベクトルｈは、基底関数φkに第２取引データｘ´iを代入した値により構成される。 That is, the vector h is composed of values obtained by substituting the second transaction data x ′ i for the basis function φ k .
また、数式（１）の右辺のＩbは、ｂ×ｂの単位行列を表し、係数λは、二次の正則化項のラグランジュ未定乗数である。 Further, I b on the right side of Expression (1) represents a b × b identity matrix, and the coefficient λ is a Lagrange undetermined multiplier of the quadratic regularization term.
なお、数式（１）によりベクトルｗの成分が負になる場合、その成分を０に置き換えてよい。このようにして、推定部１２は、第１取引データ及び第２取引データを与えられた場合に、密度比を高速に推定することができる。 In addition, when the component of the vector w becomes negative according to the equation (1), the component may be replaced with 0. In this way, the estimation unit 12 can estimate the density ratio at high speed when given the first transaction data and the second transaction data.
算出部１３は、密度比関数に基づき、第２取引データの異常度を算出する。第２取引データをｘと表し、密度比関数をｒ（ｘ）と表し、異常度をａ（ｘ）と表すとき、算出部１３は、ａ（ｘ）＝−ｌｏｇ［ｒ（ｘ）］により異常度を算出してよい。ここで、ｌｏｇは自然対数であってよい。このように異常度を算出することで、密度比関数が１に近い場合に、異常度はほぼ０であり、密度比関数が１より小さくなった場合に、異常度が正のピークを示す。すなわち、第１取引データの確率分布と第２取引データの確率分布の差が小さい場合に異常度はほぼ０であり、第１取引データの確率分布と第２取引データの確率分布の乖離が大きい場合に、異常度が正のピークを示す。 The calculation unit 13 calculates the degree of abnormality of the second transaction data based on the density ratio function. When the second transaction data is represented by x, the density ratio function is represented by r (x), and the degree of abnormality is represented by a (x), the calculation unit 13 uses a (x) = − log [r (x)]. The degree of abnormality may be calculated. Here, log may be a natural logarithm. By calculating the anomaly degree in this way, the anomaly degree is almost 0 when the density ratio function is close to 1, and the anomaly degree shows a positive peak when the density ratio function is smaller than 1. That is, when the difference between the probability distribution of the first transaction data and the probability distribution of the second transaction data is small, the degree of abnormality is almost 0, and the difference between the probability distribution of the first transaction data and the probability distribution of the second transaction data is large. In this case, the abnormality degree shows a positive peak.
第１判定部１４は、異常度に基づき、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する。より具体的には、第１判定部１４は、第２取引データに含まれるデータのうち、異常度が上位所定割合に属するデータを第１被疑データと判定してよい。このように判定することで、異常度について絶対的な閾値を設定することなく、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することができる。 The first determination unit 14 determines, based on the degree of abnormality, whether the second transaction data includes first suspected data suspected of being unfair. More specifically, the first determination unit 14 may determine, as the first suspected data, the data included in the second transaction data and having an abnormal degree belonging to a higher predetermined ratio. With this determination, it is possible to determine whether the second transaction data includes the first suspected data suspected of being unfair, without setting an absolute threshold for the abnormality degree.
第１判定部１４は、第２取引データに含まれるデータのうち、異常度が上位所定割合に属し、所定期間の値幅が第１閾値以上であるデータを第１被疑データと判定してもよい。このように、異常度について絶対的な閾値を設定せずに、値幅について第１閾値を設定することで、第２取引データに不公正取引が疑われる第１被疑データが含まれるかより高精度に判定することができる。 Of the data included in the second transaction data, the first determination unit 14 may determine that data in which the degree of abnormality belongs to the upper predetermined ratio and the price range of the predetermined period is equal to or greater than the first threshold is the first suspected data. .. As described above, by setting the first threshold value for the price range without setting the absolute threshold value for the abnormality degree, it is possible to determine whether the second transaction data includes the first suspected data that is suspected of being unfair. Can be determined.
また、第１判定部１４は、第２取引データに含まれるデータのうち、異常度が上位所定割合に属し、所定期間の出来高が第２閾値以上であるデータを第１被疑データと判定してもよい。このように、異常度について絶対的な閾値を設定せずに、出来高について第２閾値を設定することで、第２取引データに不公正取引が疑われる第１被疑データが含まれるかより高精度に判定することができる。 Further, the first determination unit 14 determines, as the first suspected data, data in which the degree of abnormality belongs to the upper predetermined ratio of the data included in the second transaction data and the volume of sales for the predetermined period is equal to or more than the second threshold value. Good. In this way, by setting the second threshold value for the volume without setting the absolute threshold value for the abnormality degree, it is possible to determine whether the second transaction data includes the first suspected data that is suspected of being unfair. Can be determined.
第２判定部１５は、予め定められた基準に基づき、第２取引データに不公正取引が疑われる第２被疑データが含まれるか判定する。第２判定部１５は、例えば、所定期間の売買注文量が閾値以上であるか、所定期間の売買注文のキャンセル量が閾値以上であるか、所定期間の値幅が閾値以上であるか、所定期間の出来高が閾値以上であるかといった基準を用いて、第２取引データに不公正取引が疑われる第２被疑データが含まれるか判定してよい。 The second determination unit 15 determines whether the second transaction data includes second suspected data suspected of unfair transaction, based on a predetermined standard. For example, the second determination unit 15 determines whether the sales order quantity in the predetermined period is equal to or larger than a threshold value, the cancellation quantity of the sales order in the predetermined period is equal to or larger than the threshold value, the price range in the predetermined period is equal to or larger than the threshold value, or the predetermined period. It may be determined whether the second transaction data includes the second suspected data suspected of being unfair transaction, using a criterion such as whether the volume of the transaction is equal to or more than a threshold value.
明細表生成部１６は、第２取引データの時間変化を示し、第２取引データのうち第１被疑データを識別可能に示す取引明細表を生成する。取引明細表は、第１ユーザ端末２０及び第２ユーザ端末３０に送信されてよい。取引明細表は、例えば、株式の取引データの時間変化を表すものであってよく、市場、銘柄の識別情報、顧客の識別情報、注文の識別情報、取引日時、出来高及び約定価格を含んでよいが、その他の情報を含んでもよい。取引明細表は、特定の銘柄についての約定の履歴を表すいわゆる歩み値を示す表であってよい。明細表生成部１６は、第２取引データのうち第１被疑データをハイライト表示したり、大きく表示したりすることで、強調して表示してよい。第２取引データのうち第１被疑データを識別可能に示す取引明細表によって、不公正取引が疑われる箇所が一目で把握できるようになる。 The statement table generation unit 16 shows a time change of the second transaction data, and generates a transaction statement table showing the first suspected data of the second transaction data in a distinguishable manner. The transaction details table may be transmitted to the first user terminal 20 and the second user terminal 30. The transaction details table may represent, for example, a time change of stock transaction data, and may include a market, stock identification information, customer identification information, order identification information, transaction date and time, trading volume and contract price. However, it may include other information. The transaction schedule may be a table showing a so-called trade price that represents a history of executions of a specific stock. The schedule generation unit 16 may highlight and display the first suspected data out of the second transaction data so as to emphasize the first suspected data. The transaction details table that can identify the first suspected data in the second transaction data makes it possible to grasp at a glance the location where the unfair transaction is suspected.
取引明細表は、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すものであってもよい。取引明細表は、例えば、第１被疑データを第１の色（例えばオレンジ）でハイライト表示し、第２被疑データを第２の色（例えば黄）でハイライト表示し、第１被疑データと第２被疑データの重なりを第３の色（例えば赤）でハイライト表示するものであってよい。第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、学習モデルによって抽出された第１被疑データと、予め定められた基準に基づき抽出された第２被疑データとの共通点と相違点を一目で把握することができる。 The transaction details table may show the first suspect data, the second suspect data, and the overlap between the first suspect data and the second suspect data in a distinguishable manner. In the transaction statement, for example, the first suspected data is highlighted in a first color (eg, orange), the second suspected data is highlighted in a second color (eg, yellow), and the first suspected data is displayed. The overlap of the second suspected data may be highlighted with a third color (for example, red). The first suspected data, the second suspected data, and the first suspected data extracted by the learning model by distinguishably showing the overlap between the first suspected data and the second suspected data are extracted based on a predetermined criterion. The common points and differences with the second suspected data can be understood at a glance.
グラフ生成部１７は、異常度の時間変化を示す第１グラフを生成する。第１グラフは、第１ユーザ端末２０及び第２ユーザ端末３０に送信されてよい。第１グラフは、例えば１秒毎に第２取引データの異常度をプロットしたグラフであってよい。第１グラフにより、異常度の時間変化が直感的に把握できるようになり、不公正取引が疑われる期間が一目で把握できるようになる。 The graph generation unit 17 generates a first graph showing the change over time in the degree of abnormality. The first graph may be transmitted to the first user terminal 20 and the second user terminal 30. The first graph may be a graph in which the degree of abnormality of the second transaction data is plotted every second, for example. With the first graph, it is possible to intuitively grasp the time change of the abnormality degree, and it is possible to grasp at a glance the period in which unfair trade is suspected.
グラフ生成部１７は、第２取引データの時間変化を示し、第２取引データのうち第１被疑データを識別可能に示す第２グラフをさらに生成してよい。第２グラフは、例えば株式の価格変動を示すグラフであってよく、第１被疑データに相当する区間をハイライト表示するものであってよい。第２取引データのうち第１被疑データを識別可能に示す第２グラフによって、不公正取引が疑われる箇所が一目で把握できるようになる。 The graph generation unit 17 may further generate a second graph showing the time change of the second transaction data and identifiable the first suspected data in the second transaction data. The second graph may be, for example, a graph showing stock price fluctuations, and may highlight a section corresponding to the first suspected data. Of the second transaction data, the second graph showing the first suspected data in an identifiable manner makes it possible to understand at a glance where the unfair transaction is suspected.
第２グラフは、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すものであってもよい。第２グラフは、例えば、第１被疑データの区間を第１の色（例えばオレンジ）でハイライト表示し、第２被疑データの区間を第２の色（例えば黄）でハイライト表示し、第１被疑データと第２被疑データが重なる区間を第３の色（例えば赤）でハイライト表示するものであってよい。第２グラフは、取引明細表と対応する態様で、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すものであってよい。第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、学習モデルによって抽出された第１被疑データと、予め定められた基準に基づき抽出された第２被疑データとの共通点と相違点を一目で把握することができる。 The second graph may show the first suspected data, the second suspected data, and the overlap between the first suspected data and the second suspected data in a distinguishable manner. The second graph, for example, highlights a section of the first suspected data in a first color (eg, orange) and highlights a section of the second suspected data in a second color (eg, yellow). A section in which the first suspected data and the second suspected data overlap may be highlighted in a third color (for example, red). The second graph may show the first suspected data, the second suspected data, and the overlap between the first suspected data and the second suspected data in a manner corresponding to the transaction schedule so that they can be identified. The first suspected data, the second suspected data, and the first suspected data extracted by the learning model by distinguishably showing the overlap between the first suspected data and the second suspected data are extracted based on a predetermined criterion. The common points and differences with the second suspected data can be understood at a glance.
図３は、本実施形態に係る売買審査装置１０の物理的構成を示す図である。売買審査装置１０は、演算部に相当するＣＰＵ（Central Processing Unit）１０ａと、記憶部に相当するＲＡＭ（Random Access Memory）１０ｂと、記憶部に相当するＲＯＭ（Read Only Memory）１０ｃと、通信部１０ｄと、入力部１０ｅと、表示部１０ｆと、を有する。これらの各構成は、バスを介して相互にデータ送受信可能に接続される。なお、本例では売買審査装置１０が一台のコンピュータで構成される場合について説明するが、売買審査装置１０は、複数のコンピュータが組み合わされて実現されてもよい。また、図３で示す構成は一例であり、売買審査装置１０はこれら以外の構成を有してもよいし、これらの構成のうち一部を有さなくてもよい。 FIG. 3 is a diagram showing a physical configuration of the trading examination device 10 according to the present embodiment. The transaction inspection device 10 includes a CPU (Central Processing Unit) 10a corresponding to a calculation unit, a RAM (Random Access Memory) 10b corresponding to a storage unit, a ROM (Read Only Memory) 10c corresponding to a storage unit, and a communication unit. It has 10d, an input unit 10e, and a display unit 10f. These respective configurations are connected to each other via a bus so that data can be transmitted and received. In this example, the case where the trading examination device 10 is composed of one computer will be described, but the trading examination device 10 may be realized by combining a plurality of computers. Moreover, the configuration shown in FIG. 3 is an example, and the trading inspection device 10 may have a configuration other than these, or may not have some of these configurations.
ＣＰＵ１０ａは、ＲＡＭ１０ｂ又はＲＯＭ１０ｃに記憶されたプログラムの実行に関する制御やデータの演算、加工を行う制御部である。ＣＰＵ１０ａは、取引データに不公正取引が疑われるデータが含まれるか判定するプログラム（売買審査プログラム）を実行する演算部である。ＣＰＵ１０ａは、入力部１０ｅや通信部１０ｄから種々のデータを受け取り、データの演算結果を表示部１０ｆに表示したり、ＲＡＭ１０ｂやＲＯＭ１０ｃに格納したりする。 The CPU 10a is a control unit that controls the execution of a program stored in the RAM 10b or the ROM 10c, calculates data, and processes the data. The CPU 10a is an arithmetic unit that executes a program (trading examination program) for determining whether or not the transaction data includes data suspected of being unfair. The CPU 10a receives various data from the input unit 10e and the communication unit 10d, displays the calculation result of the data on the display unit 10f, and stores it in the RAM 10b and the ROM 10c.
ＲＡＭ１０ｂは、記憶部のうちデータの書き換えが可能なものであり、例えば半導体記憶素子で構成されてよい。ＲＡＭ１０ｂは、ＣＰＵ１０ａが実行する売買審査プログラム、取引データといったデータを記憶してよい。なお、これらは例示であって、ＲＡＭ１０ｂには、これら以外のデータが記憶されていてもよいし、これらの一部が記憶されていなくてもよい。 The RAM 10b is a storage unit in which data can be rewritten, and may be composed of, for example, a semiconductor storage element. The RAM 10b may store data such as a trading examination program executed by the CPU 10a and transaction data. Note that these are merely examples, and data other than these may be stored in the RAM 10b, or some of these may not be stored.
ＲＯＭ１０ｃは、記憶部のうちデータの読み出しが可能なものであり、例えば半導体記憶素子で構成されてよい。ＲＯＭ１０ｃは、例えば売買審査プログラムや、書き換えが行われないデータを記憶してよい。 The ROM 10c is a storage unit capable of reading data, and may be configured by, for example, a semiconductor storage element. The ROM 10c may store, for example, a trading examination program or data that is not rewritten.
通信部１０ｄは、売買審査装置１０を他の機器に接続するインターフェースである。通信部１０ｄは、インターネット等の通信ネットワークＮに接続されてよい。 The communication unit 10d is an interface that connects the trade examination device 10 to another device. The communication unit 10d may be connected to a communication network N such as the Internet.
入力部１０ｅは、ユーザからデータの入力を受け付けるものであり、例えば、キーボード及びタッチパネルを含んでよい。 The input unit 10e receives data input from the user, and may include, for example, a keyboard and a touch panel.
表示部１０ｆは、ＣＰＵ１０ａによる演算結果を視覚的に表示するものであり、例えば、ＬＣＤ（Liquid Crystal Display）により構成されてよい。表示部１０ｆは、取引データの取引明細表と、異常度の第１グラフと、取引データの第２グラフとを表示してよい。 The display unit 10f visually displays the calculation result by the CPU 10a, and may be configured by, for example, an LCD (Liquid Crystal Display). The display unit 10f may display a transaction details table of transaction data, a first graph of the degree of abnormality, and a second graph of transaction data.
売買審査プログラムは、ＲＡＭ１０ｂやＲＯＭ１０ｃ等のコンピュータによって読み取り可能な記憶媒体に記憶されて提供されてもよいし、通信部１０ｄにより接続される通信ネットワークを介して提供されてもよい。売買審査装置１０では、ＣＰＵ１０ａが売買審査プログラムを実行することにより、図２を用いて説明した様々な動作が実現される。なお、これらの物理的な構成は例示であって、必ずしも独立した構成でなくてもよい。例えば、売買審査装置１０は、ＣＰＵ１０ａとＲＡＭ１０ｂやＲＯＭ１０ｃが一体化したＬＳＩ（Large-Scale Integration）を備えていてもよい。 The trade examination program may be provided by being stored in a computer-readable storage medium such as the RAM 10b or the ROM 10c, or may be provided via a communication network connected by the communication unit 10d. In the trading examination device 10, the CPU 10a executes the trading examination program, whereby various operations described with reference to FIG. 2 are realized. Note that these physical configurations are mere examples and do not necessarily have to be independent configurations. For example, the trade inspection device 10 may include an LSI (Large-Scale Integration) in which the CPU 10a and the RAM 10b and the ROM 10c are integrated.
図４は、本実施形態に係る売買審査装置１０により実行される第１処理のフローチャートである。はじめに、売買審査装置１０は、１週間分の第１取引データを取得し（Ｓ１０）、それに続く１日分の第２取引データを取得する（Ｓ１１）。 FIG. 4 is a flowchart of the first process executed by the trade inspection device 10 according to the present embodiment. First, the transaction inspection device 10 acquires the first transaction data for one week (S10), and then acquires the second transaction data for one day subsequent thereto (S11).
その後、売買審査装置１０は、第１取引データと第２取引データに基づいて、密度比関数により第１取引データの確率分布と第２取引データの確率分布との密度比を推定する（Ｓ１２）。また、売買審査装置１０は、密度比関数により得られた密度比に基づき、第２取引データの異常度を算出する（Ｓ１３）。 Then, the transaction inspection device 10 estimates the density ratio between the probability distribution of the first transaction data and the probability distribution of the second transaction data by the density ratio function based on the first transaction data and the second transaction data (S12). .. Further, the transaction inspection device 10 calculates the abnormality degree of the second transaction data based on the density ratio obtained by the density ratio function (S13).
さらに、売買審査装置１０は、異常度に基づき、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する（Ｓ１４）。このように、本実施形態に係る売買審査装置１０によれば、変量に対して確率分布を仮定せずに第１取引データの確率分布と第２取引データの確率分布との密度比を推定することで、第１取引データの確率分布と第２取引データの確率分布とを直接推定する必要がなくなり、教師データを必要としない学習モデルによって第２取引データの異常度を算出することができる。これにより、教師データを必要としない学習モデルによって、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することができる。 Further, the trade inspection device 10 determines whether the second transaction data includes the first suspected data that is suspected of being unfair, based on the degree of abnormality (S14). As described above, according to the trading inspection device 10 according to the present embodiment, the density ratio between the probability distribution of the first transaction data and the probability distribution of the second transaction data is estimated without assuming the probability distribution with respect to the variables. As a result, it is not necessary to directly estimate the probability distribution of the first transaction data and the probability distribution of the second transaction data, and the degree of abnormality of the second transaction data can be calculated by a learning model that does not require teacher data. Accordingly, it is possible to determine whether or not the second transaction data includes the first suspected data suspected of being unfair transaction by the learning model that does not require the teacher data.
図５は、本実施形態に係る売買審査装置１０により実行される第１被疑データの判定処理のフローチャートである。同図では、図４に示した、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する処理（Ｓ１４）の詳細を示している。 FIG. 5 is a flowchart of the determination process of the first suspected data executed by the trading examination device 10 according to the present embodiment. The figure shows the details of the process (S14) of determining whether the second transaction data shown in FIG. 4 includes the first suspected data suspected of being unfair.
はじめに、売買審査装置１０は、異常度の上位５０％に属するデータを抽出する（Ｓ１４１）。もっとも、異常度が上位何割に属するデータを抽出するかは、適宜調整することができる。 First, the transaction inspection device 10 extracts data belonging to the top 50% of abnormalities (S141). However, it is possible to appropriately adjust the percentage of the data with the highest degree of abnormality to be extracted.
売買審査装置１０は、異常度が上位５０％に属するデータのうち、値幅が第１閾値以上のデータに絞り込みを行う（Ｓ１４２）。さらに、売買審査装置１０は、出来高が第２閾値以上のデータに絞り込みを行う（Ｓ１４３）。そして、売買審査装置１０は、絞り込まれたデータを第１被疑データと判定する（Ｓ１４４）。 The transaction inspection device 10 narrows down the data whose price range is equal to or higher than the first threshold value among the data items having the highest degree of abnormality of 50% (S142). Further, the transaction inspection device 10 narrows down the data to the data of which the volume is equal to or more than the second threshold value (S143). Then, the transaction inspection device 10 determines that the narrowed-down data is the first suspected data (S144).
図６は、本実施形態に係る売買審査装置１０により実行される第２処理のフローチャートである。同図では、図４に示した第１処理の後に行われる第２処理を示している。 FIG. 6 is a flowchart of the second process executed by the trading inspection device 10 according to this embodiment. In the same figure, the 2nd process performed after the 1st process shown in FIG. 4 is shown.
はじめに、売買審査装置１０は、異常度の時間変化を示す第１グラフを生成する（Ｓ１５）。また、売買審査装置１０は、予め定められた基準に基づき、第２取引データに不公正取引が疑われる第２被疑データが含まれるか判定する（Ｓ１６）。 First, the trading inspection device 10 generates a first graph showing a temporal change in the degree of abnormality (S15). Further, the transaction inspection device 10 determines whether the second transaction data includes second suspected data suspected of being unfair transaction based on a predetermined standard (S16).
そして、売買審査装置１０は、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示す取引明細表を生成する（Ｓ１７）。また、売買審査装置１０は、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すグラフを生成する（Ｓ１８）。 Then, the transaction inspection device 10 generates a transaction statement table that can identify the first suspected data, the second suspected data, and the overlap between the first suspected data and the second suspected data (S17). Further, the transaction inspection device 10 generates a graph showing the first suspected data, the second suspected data, and the overlapping of the first suspected data and the second suspected data in a distinguishable manner (S18).
最後に、売買審査装置１０は、生成した第１グラフ、第２グラフ及び取引明細表をユーザ端末に送信する（Ｓ１９）。以上により、第２処理が終了する。 Finally, the transaction inspection device 10 transmits the generated first graph, second graph and transaction details table to the user terminal (S19). With the above, the second processing ends.
図７は、本実施形態に係る売買審査装置１０による取引データの分類を示す図である。売買審査装置１０は、第１判定部１４によって、第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定し、第２判定部１５によって、第２取引データに不公正取引が疑われる第２被疑データが含まれるか判定する。同図では、第１判定部１４によって不公正取引の疑いありと判定され、かつ、第２判定部１５によって不公正取引の疑いありと判定された場合を（１）と表している。このように、第１被疑データかつ第２被疑データである場合を第１種被疑データと呼ぶ。 FIG. 7 is a diagram showing classification of transaction data by the trading examination device 10 according to the present embodiment. In the transaction inspection device 10, the first determination unit 14 determines whether or not the second transaction data includes first suspected data that is suspected of being unfair, and the second determination unit 15 causes the second transaction data to be unfair. It is determined whether the second suspected data that is suspected is included. In the figure, the case where the first determination unit 14 determines that there is a suspicion of an unfair transaction and the second determination unit 15 determines that there is a suspicion of an unfair transaction is represented as (1). In this way, the case of the first suspected data and the second suspected data is referred to as type 1 suspected data.
また、第１判定部１４によって不公正取引の疑いなしと判定されたが、第２判定部１５によって不公正取引の疑いありと判定された場合を（２）と表している。このように、第１被疑データではないが第２被疑データである場合を第２種被疑データと呼ぶ。 Further, the case where the first determination unit 14 determines that there is no suspicion of unfair transaction but the second determination unit 15 determines that there is a suspicion of unfair transaction is represented as (2). In this way, the case where the data is not the first suspected data but the second suspected data is referred to as second-type suspected data.
また、第１判定部１４によって不公正取引の疑いありと判定されたが、第２判定部１５によって不公正取引の疑いなしと判定された場合を（３）と表している。このように、第１被疑データであるが第２被疑データでない場合を第３種被疑データと呼ぶ。 Further, the case where the first determination unit 14 determines that there is a suspicion of unfair transaction but the second determination unit 15 determines that there is no suspicion of an unfair transaction is represented as (3). In this way, the case of the first suspected data but not the second suspected data is referred to as type 3 suspected data.
最後に、第１判定部１４によって不公正取引の疑いなしと判定され、かつ、第２判定部１５によって不公正取引の疑いなしと判定された場合を（４）と表している。 Finally, the case where the first determination unit 14 determines that there is no suspicion of unfair transaction and the second determination unit 15 determines that there is no suspicion of unfair transaction is represented as (4).
以下に示す表１は、第２判定部１５により第２取引データに不公正取引が疑われるデータが含まれるか判定した結果と、同じ第２取引データを専門家が分析し、不公正取引が疑われるデータが含まれるか判定した結果とをまとめた表である。 In Table 1 below, an expert analyzes the same second transaction data as the result of determination by the second determination unit 15 whether the second transaction data includes data suspected of being unfair transaction, and 9 is a table summarizing the results of determining whether or not suspicious data is included.
表１によれば、ルールベースの第２判定部１５による判定では、専門家が不公正取引の疑なしと判断した９１件中８４件が、不公正取引が疑われるデータとして抽出されている。 According to Table 1, in the determination by the rule-based second determination unit 15, 84 out of 91 cases in which the expert determines that there is no suspicion of unfair trade is extracted as data in which unfair trade is suspected.
反対に、ルールベースの第２判定部１５による判定では、専門家が不公正取引の疑いありと判断した２７件中２５件が、不公正取引が疑われるデータとして抽出されている。 On the contrary, in the determination by the rule-based second determination unit 15, 25 out of 27 cases in which the expert determines that there is a suspicion of unfair trade is extracted as data in which unfair trade is suspected.
これらの数値から、第２判定部１５による判定について、不公正取引が疑われるデータの取り逃がしがないか表すRecall Rateは、（２５）／（２５＋２）×１００＝９２．５９％と高いが、不公正取引が疑われるデータの取り間違いがないか表すPrecision Rateは、（２５）／（２５＋８４）×１００＝２２．９４％と低いことがわかる。すなわち、ルールベースの第２判定部１５による判定は、取り逃しをほとんどしないが、取り間違いが多いことを意味している。そのため、従来の売買審査業務では、不公正取引が疑われるデータの中から不公正取引の疑いがないデータを取り除く作業に多大な時間が費やされている。 From these numerical values, regarding the determination by the second determination unit 15, the Recall Rate, which indicates whether or not the data suspected of being unfair transaction is missed, is as high as (25) / (25 + 2) × 100 = 92.59%. It can be seen that the Precision Rate, which indicates whether there is a mistake in the data for which fair trade is suspected, is as low as (25) / (25 + 84) × 100 = 22.94%. That is, the rule-based second determination unit 15 almost never misses the decision, but it means that there are many mistakes. Therefore, in the conventional trade inspection business, a great deal of time is spent to remove data that is not suspected of being unfair from the data that is suspected of being unfair.
次に、以下に示す表２は、第１判定部１４により、異常度に基づいて第２取引データに不公正取引が疑われるデータが含まれるか判定した結果と、同じ第２取引データを専門家が分析し、不公正取引が疑われるデータが含まれるか判定した結果とをまとめた表である。 Next, Table 2 below shows the result of determining whether the second transaction data contains data suspected of being unfair based on the degree of abnormality by the first determination unit 14 It is a table that summarizes the results of the analysis by the house and the determination of whether or not the data include suspected unfair trade.
表２によれば、異常度に基づく第１判定部１４による判定では、専門家が不公正取引の疑なしと判断した９１件中３７件が、不公正取引が疑われるデータとして抽出されている。 According to Table 2, in the judgment by the first judgment unit 14 based on the degree of abnormality, 37 out of 91 cases in which the expert judges that there is no suspicion of unfair trade is extracted as data in which unfair trade is suspected. ..
反対に、異常度に基づく第１判定部１４による判定では、専門家が不公正取引の疑いありと判断した２７件中２２件が、不公正取引が疑われるデータとして抽出されている。 On the contrary, in the determination by the first determination unit 14 based on the abnormality degree, 22 out of 27 cases in which the expert determines that there is a suspicion of unfair trade is extracted as data in which unfair trade is suspected.
これらの数値から、異常度に基づく第１判定部１４による判定について、不公正取引が疑われるデータの取り逃がしがないか表すRecall Rateは、（２２）／（２２＋５）×１００＝８１．４８％と高く、不公正取引が疑われるデータの取り間違いがないか表すPrecision Rateも、（２２）／（２２＋３７）×１００＝３７．２９％と高いことがわかる。すなわち、異常度に基づく第１判定部１４による判定は、取り逃しをほとんどせず、ルールベースの場合に比較して取り間違いも少なく、より実用に適していることを意味している。 From these numerical values, in the determination by the first determination unit 14 based on the degree of abnormality, the Recall Rate, which indicates whether or not data suspected of being unfair transaction is missed, is (22) / (22 + 5) × 100 = 81.48%. It can be seen that the Precision Rate, which is high and indicates whether there is an error in the data that is suspected of being unfair, is also high at (22) / (22 + 37) × 100 = 37.29%. In other words, the determination by the first determination unit 14 based on the degree of abnormality does not cause any omission, and there are few mistakes in taking as compared with the case of the rule base, which means that it is more suitable for practical use.
次に、以下に示す表３は、第１判定部１４により、異常度、値幅及び出来高に基づいて第２取引データに不公正取引が疑われるデータが含まれるか判定した結果と、同じ第２取引データを専門家が分析し、不公正取引が疑われるデータが含まれるか判定した結果とをまとめた表である。 Next, Table 3 below shows the same result as the result of the first determination unit 14 determining whether the second transaction data includes data suspected of being unfair based on the abnormality degree, the price range, and the volume. It is a table that summarizes the results of the analysis of transaction data by experts and the determination of whether or not data that is suspected of being unfair is included.
表３によれば、異常度、値幅及び出来高に基づく第１判定部１４による判定では、専門家が不公正取引の疑なしと判断した９１件中１７件が、不公正取引が疑われるデータとして抽出されている。 According to Table 3, in the judgment by the first judging unit 14 based on the degree of abnormality, the price range, and the volume, 17 out of 91 cases in which the expert judged that there was no suspicion of unfair trade were data suspected of being unfair trade. It has been extracted.
反対に、異常度、値幅及び出来高に基づく第１判定部１４による判定では、専門家が不公正取引の疑いありと判断した２７件中２２件が、不公正取引が疑われるデータとして抽出されている。 On the contrary, in the judgment by the first judging unit 14 based on the degree of abnormality, the price range, and the trading volume, 22 out of 27 cases in which the expert judges that there is a suspicion of unfair trade are extracted as data in which unfair trade is suspected. There is.
これらの数値から、異常度、値幅及び出来高に基づく第１判定部１４による判定について、不公正取引が疑われるデータの取り逃がしがないか表すRecall Rateは、（２２）／（２２＋５）×１００＝８１．４８％と高く、不公正取引が疑われるデータの取り間違いがないか表すPrecision Rateも、（２２）／（２２＋１７）×１００＝５６．４１％と高いことがわかる。すなわち、異常度、値幅及び出来高に基づく第１判定部１４による判定は、取り逃しをほとんどせず、取り間違いもルールベースの場合に比べて２．４６倍（５６．４１％／２２．９４％＝２．４６）少ないことを意味している。このように、本実施形態に売買審査装置１０によれば、不公正取引が疑われるデータの取り逃がしを少なくしつつ、取り間違いも少なくして、売買審査業務の効率化に資することができる。 From these numerical values, regarding the determination by the first determination unit 14 based on the degree of abnormality, the price range, and the volume, the Recall Rate indicating whether or not the data suspected of being unfair transaction is missed is (22) / (22 + 5) × 100 = 81. It is as high as 0.48%, and it can be seen that the Precision Rate, which indicates whether there is an error in the data suspected of being unfair, is also high at (22) / (22 + 17) × 100 = 56.41%. That is, the determination by the first determination unit 14 based on the degree of abnormality, the price range, and the volume of output rarely misses, and the error in the determination is 2.46 times (56.41% / 22.94% =) as compared with the rule-based case. 2.46) means less. As described above, according to the trading inspection device 10 of the present embodiment, it is possible to reduce the missed data of the data suspected to be unfair transaction and to reduce the mistakes of the data, thereby contributing to the efficiency of the trading inspection work.
図８は、本実施形態に係る売買審査装置１０により生成される取引明細表を示す図である。売買審査装置１０の明細表生成部１６は、第２取引データの時間変化を示し、第２取引データのうち第１被疑データを識別可能に示す取引明細表を生成する。本例の取引明細表は、株式の取引データの時間変化を表すものであり、「市場」、「銘柄」の識別情報、「顧客」の識別情報、「注文」の受付番号（識別情報）、「取引日」、「受注時間」、「売買区分」、「訂正区分」、「数量」及び「単価」を含む。具体的には、「市場」は「東証Ｍ」すなわち東証マザーズであり、「銘柄」は「ＡＡＡＡ」（証券コード）であり、「顧客」の識別情報は、「９２３４５６７８」等の８桁の番号であり、「受付番号」は「９８７６５４３２１」等の９桁の連続番号である。また、「取引日」は「２０１４／６／１７」（２０１４年６月１７日）であり、「受注時間」は「８：０１：５４」（８時１分５４秒）等である。さらに、「売買区分」は「買」又は「売」であり、「訂正区分」は「訂正」又は「取消」である。「数量」は、売買注文の数量を表し、「１００」等であり、「単価」は、指値注文の場合に「１４，２００」（１万４千２百円）等であり、成り行き注文の場合に「成行」である。 FIG. 8 is a diagram showing a transaction schedule created by the trading examination device 10 according to the present embodiment. The transaction schedule generation unit 16 of the transaction inspection device 10 generates a transaction transaction schedule showing the time change of the second transaction data and identifiable the first suspected data in the second transaction data. The transaction details table of this example represents changes over time in the transaction data of stocks, and includes "market", "brand" identification information, "customer" identification information, "order" acceptance number (identification information), It includes "transaction date", "order time", "trading category", "correction category", "quantity" and "unit price". Specifically, "market" is "TSE M" or TSE Mothers, "stock" is "AAAA" (stock code), and "customer" identification information is an 8-digit number such as "923345678". The "reception number" is a 9-digit serial number such as "987654321". The “transaction date” is “2014/6/17” (June 17, 2014), and the “ordering time” is “8:01:54” (8: 1: 54). Further, the "trading category" is "buy" or "sell", and the "correction category" is "correction" or "cancellation". The "quantity" represents the quantity of the buy / sell order, and is "100" or the like, and the "unit price" is "14,200" (14,200 yen) or the like in the case of the limit order. In some cases, it is "acting".
取引明細表は、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すものであってよい。本例では、取引明細表は、第１被疑データを破線で示す第１領域Ａにより強調表示し、第１被疑データと第２被疑データの重なりを一点鎖線で示す第２領域Ｂにより強調表示し、第２被疑データを二点鎖線で示す第３領域Ｃにより強調表示している。このように、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、学習モデルによって抽出された第１被疑データと、予め定められた基準に基づき抽出された第２被疑データとの共通点と相違点を一目で把握することができる。なお、本例では、図示を簡明にするため第１領域Ａ、第２領域Ｂ及び第３領域Ｃによって第１被疑データ、第１被疑データと第２被疑データの重なり及び第２被疑データを識別可能に示しているが、売買審査装置１０は、これらのデータをハイライト表示したり、大きく表示したりすることで、強調して表示してもよい。 The transaction details table may indicate the first suspected data, the second suspected data, and the overlap between the first suspected data and the second suspected data in a distinguishable manner. In this example, the transaction details table highlights the first suspected data by the first area A indicated by a broken line, and highlights the overlap between the first suspected data and the second suspected data by the second area B indicated by a dashed line. , The second suspected data is highlighted by a third region C indicated by a chain double-dashed line. In this way, the first suspected data, the second suspected data, and the first suspected data extracted by the learning model and the predetermined criterion are shown by distinguishably showing the overlap between the first suspected data and the second suspected data. The common points and differences with the second suspected data extracted based on can be grasped at a glance. In the present example, the first suspect data, the overlap of the first suspect data and the second suspect data, and the second suspect data are identified by the first area A, the second area B, and the third area C for the sake of simplicity. Although shown as possible, the trading examination device 10 may highlight and display these data or display them in a large size so as to be highlighted.
図９は、本実施形態に係る売買審査装置１０により生成される第１グラフを示す図である。売買審査装置１０のグラフ生成部１７は、異常度の時間変化を示す第１グラフを生成する。第１グラフは、例えば１秒毎に第２取引データの異常度をプロットしたグラフであってよい。本例では、横軸に時刻を示し、縦軸に異常度の数値を示して、買い注文の取り消し数量に関する異常度ａ１を実線で示し、売り注文の取り消し数量に関する異常度ａ２を破線で示し、買い気配数量に関する異常度ａ３を一点鎖線で示し、売り気配数量に関する異常度ａ４を二点鎖線で示している。 FIG. 9 is a diagram showing a first graph generated by the trading examination device 10 according to the present embodiment. The graph generation unit 17 of the transaction inspection device 10 generates a first graph showing a temporal change in the degree of abnormality. The first graph may be a graph in which the degree of abnormality of the second transaction data is plotted every second, for example. In this example, the horizontal axis represents time, and the vertical axis represents the numerical value of the abnormality degree, the abnormality degree a1 related to the cancel quantity of the buy order is indicated by a solid line, and the abnormality degree a2 related to the canceled quantity of the sell order is indicated by a broken line. The anomaly degree a3 relating to the buy quote quantity is indicated by a one-dot chain line, and the abnormality degree a4 relating to the sell quote quantity is indicated by a two-dot chain line.
本例の第１グラフによれば、９時頃に買い気配数量に関する異常度ａ３及び売り気配数量に関する異常度ａ４が二本のピークを形成している。また、１０時頃から１０時半頃に４種類の異常度全てが複数のピークを形成している。さらに、１３時半頃に売り気配数量に関する異常度ａ４がピークを形成しており、１４時半頃に買い気配数量に関する異常度ａ３がピークを形成している。このように、第１グラフにより、異常度の時間変化が直感的に把握できるようになり、不公正取引が疑われる期間が一目で把握できるようになる。 According to the first graph of this example, two peaks are formed at around 9 o'clock, the abnormality degree a3 relating to the buy quote quantity and the abnormality degree a4 relating to the sell quote quantity. Further, all of the four types of abnormalities form a plurality of peaks from about 10:00 to about 10:30. Further, the abnormality degree a4 regarding the sell quote amount peaks at around 13:30, and the abnormality degree a3 regarding the buy quote amount peaks around 14:30. In this way, the first graph makes it possible to intuitively grasp the temporal change in the degree of abnormality, and it becomes possible to grasp at a glance the period in which unfair trade is suspected.
図１０は、本実施形態に係る売買審査装置１０により生成される第２グラフを示す図である。売買審査装置１０のグラフ生成部１７は、第２取引データの時間変化を示し、第２取引データのうち第１被疑データを識別可能に示す第２グラフをさらに生成してよい。本例の第２グラフは、横軸に時刻を示し、縦軸に種々の取引データの値を示している。具体的には、上段から順に、買い気配数量と売り気配数量（ｑｕｏｔｅ ｖｏｌｕｍｅ）、買い注文の取り消し数量と売り注文の取り消し数量（ｃａｎｃｅｌ ｖｏｌｕｍｅ）、買い注文の約定待ち数量に対する取り消し数量の比と売り注文の約定待ち数量に対する取り消し数量の比（ｃａｎｃｅｌ ｖｏｌｕｍｅ ｒａｔｉｏ ｏｖｅｒ ｏｒｄｅｒ ｂｏｏｋ ｖｏｌｕｍｅ）、買い注文の約定待ち上位１０注文の数量と売り注文の約定待ち上位１０注文の数量（ｏｒｄｅｒ ｂｏｏｋ ｖｏｌｕｍｅ ｏｎ １０ ｌｅｖｅｌｓ）、出来高（ｔｒａｄｅｄ ｖｏｌｕｍｅ）及び価格推移（ｐｒｉｃｅ ｔｒａｎｓｉｔｉｏｎ）を含む。 FIG. 10 is a diagram showing a second graph generated by the trading examination device 10 according to the present embodiment. The graph generation unit 17 of the trade inspection device 10 may further generate a second graph showing the time change of the second transaction data and identifiable the first suspected data in the second transaction data. In the second graph of this example, the horizontal axis represents time and the vertical axis represents values of various transaction data. Specifically, in the order from the top, the bid and sell quotes (quote volume), the cancel quantity of the buy order and the cancel quantity of the sell order (cancel volume), and the ratio of the cancel quantity to the working quantity of the buy order and the sell quantity. Ratio of cancel volume to working quantity of order (cancel volume ratio over order book volume), quantity of top 10 working order of buy order and quantity of top 10 working order of sell order (order book volume on 10 levels), Includes traded volume and price transition.
本例では、第２グラフは、第１被疑データを破線で示す第１領域Ａにより強調表示し、第１被疑データと第２被疑データの重なりを一点鎖線で示す第２領域Ｂにより強調表示し、第２被疑データを二点鎖線で示す第３領域Ｃにより強調表示している。このように、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、学習モデルによって抽出された第１被疑データと、予め定められた基準に基づき抽出された第２被疑データとの共通点と相違点を一目で把握することができる。なお、本例では、図示を簡明にするため第１領域Ａ、第２領域Ｂ及び第３領域Ｃによって第１被疑データ、第１被疑データと第２被疑データの重なり及び第２被疑データを識別可能に示しているが、売買審査装置１０は、これらのデータをハイライト表示したり、大きく表示したりすることで、強調して表示してもよい。 In this example, in the second graph, the first suspected data is highlighted by the first area A indicated by a broken line, and the overlap between the first suspected data and the second suspected data is highlighted by the second area B indicated by a dashed line. , The second suspected data is highlighted by a third region C indicated by a chain double-dashed line. In this way, the first suspected data, the second suspected data, and the first suspected data extracted by the learning model and the predetermined criterion are shown by distinguishably showing the overlap between the first suspected data and the second suspected data. The common points and differences with the second suspected data extracted based on can be grasped at a glance. In the present example, the first suspect data, the overlap of the first suspect data and the second suspect data, and the second suspect data are identified by the first area A, the second area B, and the third area C for the sake of simplicity. Although shown as possible, the trading examination device 10 may highlight and display these data or display them in a large size so as to be highlighted.
第２グラフに表示されている第１領域Ａ、第２領域Ｂ及び第３領域Ｃは、図８に示す取引明細表に表示されている第１領域Ａ、第２領域Ｂ及び第３領域Ｃと対応している。このように、第２グラフ及び取引明細表の両方で第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりを識別可能に示すことで、第２グラフ及び取引明細表で情報を補完しながら被疑データの詳細を確認することができる。 The first area A, the second area B and the third area C displayed in the second graph are the first area A, the second area B and the third area C displayed in the transaction statement shown in FIG. It corresponds to. In this way, the first graph suspect data, the second suspect data, and the overlap of the first suspect data and the second suspect data are distinguishably shown in both the second graph and the transaction schedule, so that the second graph and the transaction schedule can be identified. You can check the details of the suspected data while supplementing the information with.
なお、本例では、図９に示す第１グラフと図１０に示す第２グラフとを別々に図示しているが、第１グラフと第２グラフは、横軸を共有するように並べて表示されてよい。第１グラフ及び第２グラフを、横軸を共有するように表示することで、異常度がピークを形成している時刻と、第１被疑データ、第２被疑データ及び第１被疑データと第２被疑データの重なりとを容易に比較することができる。 In this example, the first graph shown in FIG. 9 and the second graph shown in FIG. 10 are shown separately, but the first graph and the second graph are displayed side by side so that the horizontal axis is shared. You may. By displaying the first graph and the second graph so that the horizontal axis is shared, the time when the abnormality degree forms a peak, the first suspected data, the second suspected data, and the first suspected data and the second suspected data are displayed. The overlap of suspected data can be easily compared.
以上説明した実施形態は、本発明の理解を容易にするためのものであり、本発明を限定して解釈するためのものではない。実施形態が備える各要素並びにその配置、材料、条件、形状及びサイズ等は、例示したものに限定されるわけではなく適宜変更することができる。また、異なる実施形態で示した構成同士を部分的に置換し又は組み合わせることが可能である。 The embodiments described above are for facilitating the understanding of the present invention, and are not for limiting the interpretation of the present invention. Each element included in the embodiment and its arrangement, material, condition, shape, size and the like are not limited to the exemplified ones but can be changed as appropriate. Further, the configurations shown in different embodiments can be partially replaced or combined.
１０…売買審査装置、１０ａ…ＣＰＵ、１０ｂ…ＲＡＭ、１０ｃ…ＲＯＭ、１０ｄ…通信部、１０ｅ…入力部、１０ｆ…表示部、１１…取得部、１２…推定部、１３…算出部、１４…第１判定部、１５…第２判定部、１６…明細表生成部、１７…グラフ生成部、２０…第１ユーザ端末、３０…第２ユーザ端末、４０…証券会社データベース、５０…情報ベンダデータベース、１００…売買審査システム、ＤＢ１…顧客データ、ＤＢ２…取引データ、ＤＢ３…マーケットデータ 10 ... Trading examination device, 10a ... CPU, 10b ... RAM, 10c ... ROM, 10d ... Communication part, 10e ... Input part, 10f ... Display part, 11 ... Acquisition part, 12 ... Estimating part, 13 ... Calculation part, 14 ... 1st determination part, 15 ... 2nd determination part, 16 ... Detailed table generation part, 17 ... Graph generation part, 20 ... 1st user terminal, 30 ... 2nd user terminal, 40 ... Securities company database, 50 ... Information vendor database , 100 ... Trading examination system, DB1 ... Customer data, DB2 ... Transaction data, DB3 ... Market data
Claims (14)
前記密度比関数に基づき、前記第２取引データの異常度を算出する算出部と、
前記異常度に基づき、前記第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する第１判定部と、
を備える売買審査装置。 A probability distribution of the first transaction data and a second distribution of the first transaction data by a density ratio function based on first transaction data relating to the buying and selling of financial products and second transaction data relating to the trading of the financial instruments and different from the first transaction data. An estimation unit that estimates the density ratio with the probability distribution of transaction data,
A calculation unit that calculates an abnormality degree of the second transaction data based on the density ratio function;
A first determining unit that determines whether the second transaction data includes first suspected data suspected of being unfair, based on the degree of abnormality;
A trading examination device equipped with.
請求項１に記載の売買審査装置。 The first transaction data is transaction data for a longer period than the second transaction data,
The trading examination device according to claim 1.
請求項１又は２に記載の売買審査装置。 The first transaction data and the second transaction data include data relating to the sales order quantity for a predetermined period,
The trading examination device according to claim 1 or 2.
請求項１から３のいずれか一項に記載の売買審査装置。 The first transaction data and the second transaction data include data regarding the amount of cancellation of a sales order for a predetermined period,
The trade examination device according to claim 1.
請求項１から４のいずれか一項に記載の売買審査装置。 The first determination unit determines, as the first suspected data, data in which the degree of abnormality belongs to a higher predetermined ratio among the data included in the second transaction data.
The trade examination device according to claim 1.
請求項５に記載の売買審査装置。 Of the data included in the second transaction data, the first determination unit determines, as the first suspected data, data in which the degree of abnormality belongs to a higher predetermined ratio and the price range of a predetermined period is equal to or more than a first threshold value. ,
The trading inspection device according to claim 5.
請求項５又は６に記載の売買審査装置。 Of the data included in the second transaction data, the first determination unit determines that the data in which the degree of abnormality belongs to a higher predetermined ratio and the volume of sales for a predetermined period is a second threshold value or more is the first suspected data. ,
The trade examination device according to claim 5 or 6.
請求項１から７のいずれか一項に記載の売買審査装置。 Further comprising a graph generation unit that generates a first graph showing the temporal change of the abnormality degree,
The trading inspection device according to claim 1.
請求項８に記載の売買審査装置。 The graph generation unit further generates a second graph showing a time change of the second transaction data and identifiably the first suspected data in the second transaction data,
The trading inspection device according to claim 8.
前記第２グラフは、前記第１被疑データ、前記第２被疑データ及び前記第１被疑データと前記第２被疑データの重なりを識別可能に示す、
請求項９に記載の売買審査装置。 Further comprising a second determination unit for determining whether or not the second transaction data includes second suspected data suspected of being unfair, based on a predetermined standard.
The second graph shows the first suspect data, the second suspect data, and the overlap between the first suspect data and the second suspect data in a distinguishable manner.
The trading examination device according to claim 9.
請求項１から１０のいずれか一項に記載の売買審査装置。 A detailed statement generation unit is further provided that indicates a time change of the second transaction data and generates a transaction detailed statement that identifies the first suspected data among the second transaction data.
The trade examination device according to claim 1.
前記取引明細表は、前記第１被疑データ、前記第２被疑データ及び前記第１被疑データと前記第２被疑データの重なりを識別可能に示す、
請求項１１に記載の売買審査装置。 Further comprising a second determination unit for determining whether or not the second transaction data includes second suspected data suspected of being unfair, based on a predetermined standard.
The transaction details table shows the first suspect data, the second suspect data, and the overlap between the first suspect data and the second suspect data in a distinguishable manner.
The trading examination device according to claim 11.
前記密度比関数に基づき、前記第２取引データの異常度を算出することと、
前記異常度に基づき、前記第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定することと、
を含む売買審査方法。 A probability distribution of the first transaction data and a second distribution of the first transaction data by a density ratio function based on first transaction data relating to the buying and selling of financial products and second transaction data relating to the trading of the financial instruments and different from the first transaction data. Estimating the density ratio with the probability distribution of the transaction data,
Calculating an abnormality degree of the second transaction data based on the density ratio function,
Determining whether the second transaction data includes first suspect data suspected of being unfair, based on the degree of abnormality;
Trading examination method including.
金融商品の売買に関する第１取引データと、前記金融商品の売買に関し、前記第１取引データと異なる第２取引データとに基づいて、密度比関数により前記第１取引データの確率分布と前記第２取引データの確率分布との密度比を推定する推定部、
前記密度比関数に基づき、前記第２取引データの異常度を算出する算出部、及び
前記異常度に基づき、前記第２取引データに不公正取引が疑われる第１被疑データが含まれるか判定する第１判定部、
として機能させる売買審査プログラム。 A computer equipped with a transaction examination device,
A probability distribution of the first transaction data and a second distribution of the first transaction data by a density ratio function based on first transaction data relating to the buying and selling of financial products and second transaction data relating to the trading of the financial instruments and different from the first transaction data. An estimation unit that estimates the density ratio with the probability distribution of transaction data,
Based on the density ratio function, a calculator that calculates the degree of abnormality of the second transaction data, and based on the degree of abnormality, it is determined whether the second transaction data includes first suspected data suspected of unfair transaction. The first determination unit,
A trading review program to function as.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2018206232A JP6932680B2 (en) | 2018-11-01 | 2018-11-01 | Trading Examination Equipment, Trading Examination Method and Trading Examination Program |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2018206232A JP6932680B2 (en) | 2018-11-01 | 2018-11-01 | Trading Examination Equipment, Trading Examination Method and Trading Examination Program |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JP2020071751A true JP2020071751A (en) | 2020-05-07 |
| JP6932680B2 JP6932680B2 (en) | 2021-09-08 |
Family
ID=70547867
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP2018206232A Active JP6932680B2 (en) | 2018-11-01 | 2018-11-01 | Trading Examination Equipment, Trading Examination Method and Trading Examination Program |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP6932680B2 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113127894A (en) * | 2021-03-03 | 2021-07-16 | 合肥工业大学 | Public opinion propagation model-based cryptocurrency network data propagation method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002269349A (en) * | 2001-03-13 | 2002-09-20 | Artis Kk | Transaction execution system and its method, and recording medium for recording transaction execution program operated on computer |
| JP2008021141A (en) * | 2006-07-13 | 2008-01-31 | Kabu.Com Securities Co Ltd | Purchase order receiving system and purchase order receiving method |
| JP2010146093A (en) * | 2008-12-16 | 2010-07-01 | Tokyo Stock Exchange Inc | Securities transaction extraction device, securities transaction extraction method, and program thereof |
| JP2012242985A (en) * | 2011-05-18 | 2012-12-10 | Hitachi-Ge Nuclear Energy Ltd | Equipment abnormality determination device and equipment abnormality determination method |
-
2018
- 2018-11-01 JP JP2018206232A patent/JP6932680B2/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002269349A (en) * | 2001-03-13 | 2002-09-20 | Artis Kk | Transaction execution system and its method, and recording medium for recording transaction execution program operated on computer |
| JP2008021141A (en) * | 2006-07-13 | 2008-01-31 | Kabu.Com Securities Co Ltd | Purchase order receiving system and purchase order receiving method |
| JP2010146093A (en) * | 2008-12-16 | 2010-07-01 | Tokyo Stock Exchange Inc | Securities transaction extraction device, securities transaction extraction method, and program thereof |
| JP2012242985A (en) * | 2011-05-18 | 2012-12-10 | Hitachi-Ge Nuclear Energy Ltd | Equipment abnormality determination device and equipment abnormality determination method |
Non-Patent Citations (5)
| Title |
|---|
| HOXO_M: "非制約最小二乗密度比推定法 uLSIF を用いた外れ値検出", インターネット, JPN6020042758, 21 July 2016 (2016-07-21), ISSN: 0004386742 * |
| SHOHEI HIDO ほか: "Statistical Outlier Detection Using Direct Density Ratio Estimation", KNOWLEDGE AND INFORMATION SYSTEMS, vol. 26, no. 2, JPN6020042761, February 2011 (2011-02-01), pages 309 - 336, XP019873555, ISSN: 0004386745, DOI: 10.1007/s10115-010-0283-2 * |
| カブドットコム証券株式会社: "国内証券初 売買審査業務への人工知能（AI）の導入について", 企業開示情報総合サイト QUICK DISCLOSURE： HTTPS://DISCL.QUICK.CO.JP/, JPN6020042759, 29 June 2018 (2018-06-29), ISSN: 0004386743 * |
| 寺田 英雄: "異常検知 〜機械学習の有望なアプリケーション〜", QCON TOKYO 2014 セッション資料, JPN6020042757, 30 April 2014 (2014-04-30), ISSN: 0004386741 * |
| 株式会社東京商品取引所: "当社における取引内容の審査（売買審査）について", 東京商品取引所 WEBサイト： HTTPS://WWW.TOCOM.OR.JP/JP/, JPN6020042760, 12 August 2016 (2016-08-12), ISSN: 0004386744 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113127894A (en) * | 2021-03-03 | 2021-07-16 | 合肥工业大学 | Public opinion propagation model-based cryptocurrency network data propagation method and system |
| CN113127894B (en) * | 2021-03-03 | 2022-09-30 | 合肥工业大学 | Public opinion propagation model-based cryptocurrency network data propagation method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| JP6932680B2 (en) | 2021-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP2020523657A (en) | Method and system for evaluating vendors | |
| US20230070176A1 (en) | Architecture for data processing and user experience to provide decision support | |
| JP5346816B2 (en) | Algorithmic trading | |
| EP2866168B1 (en) | Calibration of strategies for fraud detection | |
| US11593305B2 (en) | Data conversion and distribution systems | |
| CN109872015A (en) | Method and device for behavioral data assessment | |
| Li et al. | Dynamic price discovery in Chinese agricultural futures markets | |
| JP2016099915A (en) | Server for credit examination, system for credit examination, and program for credit examination | |
| CN112884547A (en) | Intelligent commodity recommendation method, device, medium and terminal equipment | |
| JP6794431B2 (en) | Initial Margin Methods and systems for calculating and providing initial margin based on the Standard Model | |
| JP6932680B2 (en) | Trading Examination Equipment, Trading Examination Method and Trading Examination Program | |
| JPH11175602A (en) | Credit risk measuring device | |
| CN113744024B (en) | Merchant matching method and device, computer equipment and storage medium | |
| KR101829741B1 (en) | Server, Method, and Computer-Readable Medium for Providing Used Car Information | |
| JP2010003194A (en) | Portfolio risk information providing device, server | |
| CA3081254C (en) | Data conversion and distribution systems | |
| US20220044341A1 (en) | System and method for rapid evaluation of raw material price risk mitigation contracts | |
| Bianchi | Are multi-factor Gaussian term structure models still useful? An empirical analysis on Italian BTPs | |
| TWM557882U (en) | Warrant commodities evaluation system | |
| JP2012118612A (en) | Marketing proposal support system | |
| JP7420223B2 (en) | Unfair trade visualization device, method and program | |
| TWI769385B (en) | Method and system for screening potential purchasers of financial products | |
| Kupabado et al. | Price effects of commodity financialization: Review of the evidence | |
| WO2016141099A1 (en) | Entity state data detection unit | |
| CN112258246A (en) | Abnormal offer identification method and device for material, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A80 | Written request to apply exceptions to lack of novelty of invention |
Free format text: JAPANESE INTERMEDIATE CODE: A80 Effective date: 20181119 |
|
| A621 | Written request for application examination |
Free format text: JAPANESE INTERMEDIATE CODE: A621 Effective date: 20191129 |
|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20200923 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20201117 |
|
| A601 | Written request for extension of time |
Free format text: JAPANESE INTERMEDIATE CODE: A601 Effective date: 20210113 |
|
| A521 | Request for written amendment filed |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20210305 |
|
| TRDD | Decision of grant or rejection written | ||
| A01 | Written decision to grant a patent or to grant a registration (utility model) |
Free format text: JAPANESE INTERMEDIATE CODE: A01 Effective date: 20210730 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20210818 |
|
| R150 | Certificate of patent or registration of utility model |
Ref document number: 6932680 Country of ref document: JP Free format text: JAPANESE INTERMEDIATE CODE: R150 |