JP3692765B2 - Communication status judging device in communication network - Google Patents
Communication status judging device in communication network Download PDFInfo
- Publication number
- JP3692765B2 JP3692765B2 JP04853998A JP4853998A JP3692765B2 JP 3692765 B2 JP3692765 B2 JP 3692765B2 JP 04853998 A JP04853998 A JP 04853998A JP 4853998 A JP4853998 A JP 4853998A JP 3692765 B2 JP3692765 B2 JP 3692765B2
- Authority
- JP
- Japan
- Prior art keywords
- normal
- connection
- communication network
- point likelihood
- calculating
- 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.)
- Expired - Fee Related
Links
Images
Description
【０００１】
【発明の属する技術分野】
本発明は、通信網における疎通状況判定装置に関し、特に通信網における単位時間当りの呼の接続要求数（ＢＩＤ) と単位時間当りの呼の接続完了数（ＡＮＳ) とから接続成功率を求めて通信網における疎通状況の正常又は異常を判定する装置に関する。
【０００２】
【従来の技術】
国際電話網の疎通管理について、ＩＴＵ、ＣＣＩＴＴの勧告では、国際通信事業者は「通話の接続に成功した率であるＡＢＲ（Ａｎｓｗｅｒ Ｂｉｄ Ｒａｔｉｏ）を対外通信事業者（キャリア）別に観測し、ＡＢＲが低下した場合は回線異常やトラヒックの急増等の現象が起こっていないか調査する」ことになっている。これらの基準は、ＩＴＵ，ＣＣＩＴＴ，ＲＥＤ ＢＯＯＫ，Ｖｏｌｕｍｅ II，Ｆａｓｃｉｃｌｅ II．３（１９８５），ｐｐ．５−８１やＩＴＵ，Ｒｅｃｏｍｍｅｎｄａｔｉｏｎｓ Ｅ．４０１−Ｅ．４２７，ＢＬＵＥ ＢＯＯＫ，ＶｏｌｕｍｅII，Ｆａｓｃｉｃｌｅ II．３（１９８９）に規定されている。
【０００３】
図１は、ある国に対する１日の接続呼におけるＡＢＲ、ＢＩＤ及びＡＮＳの時系列データを示す特性図である。
【０００４】
図１（ａ）において、破線及び一点鎖線は、それぞれ２つの段階の異常を示す閾値である。これら閾値は、当該国における過去のデータに基づいて監視員の経験上から設定される。呼の接続要求数が一日の中でも多い時間帯（例えば午後５時〜午後７時）では閾値が高くなっており、これにより監視処理が円滑に行われている。しかしながら、図１（ｂ）に示すようにＢＩＤが極めて減少する時間帯（例えば午前１時〜午前７時）では、図１（ａ）に示すようにＡＢＲが激しく上下する。この現象は、ＡＢＲをＡＮＳとＢＩＤとの比から求めている（ＡＢＲ＝ＡＮＳ／ＢＩＤ）ためである。即ち、ＡＮＳ及びＢＩＤが増大する時間帯では、両者の値が共に大きいためにＡＢＲの変動は少ない。逆に、ＡＮＳ及びＢＩＤが極めて減少する時間帯では、両者の値が極めて小さいため、ＡＢＲの変動が大きくなる。
【０００５】
従って、後者の時間帯では、客観的にみて疎通状況が悪くないのにＡＢＲが頻繁に閾値を越えてしまい、監視員が重要なアラームを見つけることが困難になる。異常時に正確にアラームが発生するように閾値を設定するには、監視員の多大な経験を必要とする。また、多数の監視対象地域の各々に閾値を設定するには、かなりの人手を要する。
【０００６】
図２は、従来の判定手法における任意の時間帯における観測点と判別関数との関係を示している。この手法では、ＢＩＤ−ＡＮＳ特性の原点を通る直線で構成される閾値を監視員が経験的に設定し、ある時間帯における各観測点が閾値より下になった時を疎通状況の悪化と判定する。このように、閾値の設定が監視員の経験に依存しているので、この閾値を統計的計算に基づいて自動化することが非常に難しい。また、ＡＮＳとＢＩＤとの比から求まるＡＢＲが非線形であるため、これを直線の閾値と比較することは不自然である。さらに、観測点と閾値直線との間の距離（図中のＸ，Ｙ）がＢＩＤに応じて異なるので判定精度が一定とはならない。即ち、ＢＩＤが小さい観測点は閾値に近づいているので厳しい判定となり、逆にＢＩＤが大きい観測点は閾値から遠いので甘い判定となる。
【０００７】
このような問題点を解決する方法として、次のような判定方法が存在している。
【０００８】
対外通信事業者宛のトラヒックをさらに呼の種類毎に細分化する。呼の種類をＣ１，Ｃ２，・・・，Ｃｉ，Ｃｊ，・・・，Ｃｎ（ただし、Ｃｉ∩Ｃｊ＝０）、期間Ｔの呼の種類毎の接続要求回数をＢｉｄ１，Ｂｉｄ２，・・・，Ｂｉｄｉ，・・・，Ｂｉｄｎ、同様に期間Ｔの呼の種類毎の接続成功回数をＡｎｓ１，Ａｎｓ２，・・・，Ａｎｓｉ，・・・，Ａｎｓｎとする。ただし、呼の種類がいずれであっても次の仮定が成り立つとする。
【０００９】
この時、残差を表すｅi の分布を正規分布で近似することができ、各ｅi の確率密度関数Ｅi （ｘ）はサンプルデータの平均値と分散とから求められる。Ｂｉｄi の観測値をｂｉｄi o、Ａｎｓi の観測値をａｎｓi o、ｅi o＝（ａi √ｂｉｄi o＋ｂi ）−√ａｎｓi oとする。呼の各種類毎に得られた残差がそれぞれｅ1 o，ｅ2 o，・・・より小さくなる確率は次のようになる。
【数７】
このＰ（ｅ1 ，ｅ2 ，・・・）が監視対象の対外通信事業者毎に定めた値ηを下回った時に、網異常の可能性があると判定する。
【００１１】
【発明が解決しようとする課題】
しかしながら、このような従来の判定方法は、判定精度が十分ではなく、かつ時刻毎の判定であることからやや悪い疎通状況が長く続く場合に判定ができない恐れがあった。
【００１２】
従って本発明の目的は、判定精度が高く、やや悪い疎通状況が長く続く状態も判定できる通信網における疎通状況判定装置を提供することにある。
【００１３】
【課題を解決するための手段】
本発明によれば、通信網における単位時間当りの呼の接続要求数（ＢＩＤ）と単位時間当りの呼の接続完了数（ＡＮＳ）とから接続成功率（ＡＢＲ）を求めて通信網における疎通状況の正常又は異常を判定する装置であって、通信網が正常時のＢＩＤとＡＮＳとの組の集合を２項分布によるモデル作成のためのサンプルデータとし、サンプルデータから通信網が正常時の平均接続率を算出する２項分布を用いたモデル生成手段と、ＢＩＤ及びＡＮＳの時系列と、モデル生成手段によって生成された平均接続率とに基づいて、通信網が正常であると考える正常仮説の一点尤度と、通信網が異常であると考える異常仮説の一点尤度とを算出する一点尤度計算手段と、一点尤度計算手段により算出された正常仮説及び異常仮説の一点尤度の時系列を用いて通信網の疎通状況の正常又は異常の判定を行う判定手段とを備えた通信網における疎通状況判定装置が提供される。
【００１４】
本発明は、ＢＩＤとＡＮＳとの関係を２項分布でモデル化して、例えばＳＰＲＴ（Ｓｅｑｕｅｎｔｉａｌ Ｐｒｏｂａｂｉｌｉｔｙ Ｒａｔｉｏ Ｔｅｓｔ）法を用いて疎通状況判定を行うものである。このような判定処理を行う構成とすることにより、自動化が実現できかつ判定の精度が向上する。また、やや悪い疎通状況が長く続く状態も判定することができる。さらに、監視対象地域の状況等に柔軟に対応して一義的に定まる疎通判定のモデルを生成できる。
【００１５】
【発明の実施の形態】
図３は本発明の疎通状況判定装置の一実施形態の構成を概略的に示すブロック図である。同図において、１１はモデル生成部、１２は一点尤度計算部、１３はＳＰＲＴ判定部である。本実施形態は、ＢＩＤとＡＮＳとの関係を２項分布でモデル化してＳＰＲＴ法を用いて疎通状況判定を行い、さらにＡＩＣ基準を用いて最適な曜日群毎にモデルを生成するものである。
【００１６】
モデル生成部１１には、通信網正常時のＢＩＤとＡＮＳとの組の集合Ｓ＝｛（ｂ1 ，ａ1 ），（ｂ2 ，ａ2 ）｝が、モデル作成のためのサンプルデータとして入力される。この時、通信網正常時の平均接続率
【数８】
【数９】
一点尤度計算部１２は、網が正常と考える正常仮説及び異常と考える異常仮説の一点尤度を各時刻毎に計算する。ＢＩＤ及びＡＮＳの観測値がｂ及びａ、接続成功確率がｐであるとすると、一点尤度ｌｌ（ｂ，ａ｜ｐ）は、以下のようになる。
【数１０】
【数１１】
【数１２】
【００１８】
ＳＰＲＴ判定部１３は、ＳＰＲＴ法に基づいて正常、異常を判定する。このＳＰＲＴ法は、正常仮説と異常仮説との対数尤度の時系列を用いて正常又は異常の判定を行う方法であって、異常を見逃す確率と正常を異常と判定する確率とを一定値未満に保証した上で、判定を得るまでの観測時間を最小にできる性質を持っている。このため疎通状況の監視に適している。
【００１９】
２項分布の場合、分散は試行回数の平方根に比例する。このため、ＢＩＤ及びＡＮＳを平方根変換した後に回帰モデルを構築し、回帰直線と実際の値との残差を正規分布で近似することも考えられる。
【００２０】
下記の表１は、全対外通信事業者（キャリア）に対して７８日間分のデータを用い、従来例として平方根変換を用いたモデルと本発明の２項分布を用いたモデルとの比較している。ただし、この比較は対数尤度を用いて行っている。
【００２１】
【表１】
この表からわかるように、全ての地域において本発明の方が対数尤度が大きく、モデルとして優れていることがわかる。異常時の一点尤度ｌｌ（ｂ，ａ｜ｐw ）が正常時の一点尤度
【数１３】
【００２３】
次に、ＡＩＣを用いた曜日グループ別モデルの生成について説明する。
【００２４】
接続成功率が平日と休日とでは異なるため、曜日別にモデルを作成することでモデルの精度向上できる可能性がある。しかし、過度にモデルを作りすぎると過剰学習が起こる危険性がある。また、対外通信事業者毎に曜日の特性が異なることも考慮する必要がある。そこで、各種の曜日の組み合わせ毎にモデルを作成し、ＡＩＣ基準を用いて最適な曜日分割の方法を各対外通信事業者毎に求めてみた。
【００２５】
最適となったグループ構成を以下に示す。
｛月, 火, 水, 木, 金, 土, 日, 祝｝
｛月, 火, 水, 木, 金｝｛土, 日, 祝｝
｛月, 火, 水, 木, 金｝｛土, 日｝｛祝｝
｛月, 火, 水, 木｝｛金｝｛土, 日, 祝｝
｛月｝｛火, 水, 木, 金｝｛土, 日, 祝｝
｛月｝｛火, 水, 木, 金｝｛土, 日｝｛祝｝
｛月｝｛火, 水, 木, 金｝｛土｝｛日, 祝｝
｛月｝｛火, 水, 木｝｛金｝｛土, 日, 祝｝
｛月｝｛火, 水, 木｝｛金｝｛土｝｛日, 祝｝
｛月｝｛火, 水, 木｝｛金｝｛土, 日｝｛祝｝
｛月｝｛火, 水, 木｝｛金｝｛土｝｛日｝｛祝｝
【００２６】
以上のように、時差や文化( 例えば回教では金曜が休日) 等を反映した曜日分けができたことから、ＡＩＣ基準を用いたグループ化が有効である。
【００２７】
また、過去の７８日間分の全対外通信事業者向けのＢＩＤ及びＡＮＳの時系列に対し、異常を検出する時刻を、感度パラメータεの値を変えて求めた。
【００２８】
図５（ａ）は、種々のεに対して、異常とみなす状況を検出する回数を表わしている。εの取り得る値は０より大きく１より小さい数であり、同図はεの取り得る全範囲を網羅している。同図では、εが増えるにつれて異常検出回数が単調に増加している。従来方式の場合、閾値を対外通信事業者毎かつ曜日毎に設定しなければならないため、１つのパラメータで異常検出回数を制御することができない。これに対して本発明では、期待される異常検出回数を定めればεの値が一意に定まり、システム運用が容易になる。
【００２９】
図５（ｂ）は、異常とみなした時点において、期待接続完了数（＝ＢＩＤ×ｐ）とＡＮＳとの差を救済すべき呼の数をみなした場合の救済対象呼数とεとの関係を示している。同図はεの取り得る全範囲を網羅している。同図では、εが増えるにつれて救済対象呼の数が単調に増加している。従来方式の場合、１つのパラメータで救済呼対象呼数を制御することができない。これに対して本発明では、期待される救済対象呼数を定めればεの値が一意に定まり、システム運用が容易になる。
【００３０】
さらに、同一期間に対し、従来のシステムで異常検出を行う場合との比較を行った。従来システムの場合、異常検出回数が約５万、異常検出時の救済対象呼数は約４９万であった。一方、本発明では、ε＝０．６に設定すれば、従来システムと同等の異常検出回数を示すが、本方式の救済対象呼数は約７５万であり従来システムに比べて救済対象呼数が約２６万多い。疎通異常の検出がきっかけとなって救済対象呼を疎通させるための網制御が行われることから、検出回数がほぼ同等でも異常検出時における救済対象呼数が多い本発明の方が、従来技術より検出効果がある。
【００３１】
以上述べた実施形態は全て本発明を例示的に示すものであって限定的に示すものではなく、本発明は他の種々の変形態様及び変更態様で実施することができる。従って本発明の範囲は特許請求の範囲及びその均等範囲によってのみ規定されるものである。
【００３２】
【発明の効果】
以上説明したように本発明によれば、通信網が正常時のＢＩＤとＡＮＳとの組の集合を２項分布によるモデル作成のためのサンプルデータとし、サンプルデータから通信網が正常時の平均接続率を算出する２項分布を用いたモデル生成手段と、ＢＩＤ及びＡＮＳの時系列と、モデル生成手段によって生成された平均接続率とに基づいて、通信網が正常であると考える正常仮説の一点尤度と、通信網が異常であると考える異常仮説の一点尤度とを算出する一点尤度計算手段と、一点尤度計算手段により算出された正常仮説及び異常仮説の一点尤度の時系列を用いて通信網の疎通状況の正常又は異常の判定を行う判定手段とを備えているため、自動化が実現できかつ判定の精度が向上する。また、やや悪い疎通状況が長く続く状態も判定することができる。さらに、監視対象地域の状況等に柔軟に対応して一義的に定まる疎通判定のモデルを生成できる。
【図面の簡単な説明】
【図１】ある国に対する１日の接続呼におけるＡＢＲ、ＢＩＤ及びＡＮＳの時系列データを示す特性図である。
【図２】従来の判定手法における任意の時間帯における観測点と判別関数との関係を示す図である。
【図３】本発明の通信網における疎通状況判定装置の一実施形態の構成を概略的に示すブロック図である。
【図４】本発明における任意の時間帯における観測点と判別関数の関係を示す図である。
【図５】本発明における感度パラメータεに対する異常検出回数及び救済対象となる呼の数を示す特性図である。
【符号の説明】
１１ モデル生成部
１２ 一点尤度計算部
１３ ＳＰＲＴ判定部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a communication status determination apparatus in a communication network, and in particular, obtains a connection success rate from the number of call connection requests (BID) per unit time and the number of call connection completions (ANS) per unit time in the communication network. The present invention relates to an apparatus for determining normality or abnormality of a communication status in a communication network.
[0002]
[Prior art]
Regarding the communication management of the international telephone network, ITU and CCITT recommended that international carriers observe “ABR (Answer Bid Ratio), which is the rate of successful call connection, for each external carrier (carrier). If it falls, we will investigate whether there is a phenomenon such as a line abnormality or a sudden increase in traffic. " These standards are ITU, CCITT, RED BOOK, Volume II, Fascicle II. 3 (1985), pp. 5-81, ITU, Recommendations E.I. 401-E. 427, BLUE BOOK, Volume II, Faculty II. 3 (1989).
[0003]
FIG. 1 is a characteristic diagram showing time series data of ABR, BID, and ANS in a daily connection call to a certain country.
[0004]
In FIG. 1 (a), the broken line and the alternate long and short dash line are threshold values indicating abnormalities in two stages, respectively. These threshold values are set based on the experience of the observer based on past data in the country. The threshold value is high during a time period in which the number of call connection requests is large (for example, from 5:00 pm to 7:00 pm), whereby the monitoring process is performed smoothly. However, as shown in FIG. 1B, in the time zone in which the BID decreases extremely (for example, from 1 am to 7 am), the ABR rises and falls sharply as shown in FIG. This phenomenon is because ABR is obtained from the ratio of ANS and BID (ABR = ANS / BID). That is, in the time zone in which ANS and BID increase, since both values are large, the fluctuation of ABR is small. On the contrary, in the time zone in which ANS and BID are extremely decreased, both values are extremely small, so that the fluctuation of ABR becomes large.
[0005]
Therefore, in the latter time zone, the ABR frequently exceeds the threshold even though the communication situation is not bad from an objective viewpoint, and it becomes difficult for the monitor to find an important alarm. In order to set the threshold value so that an alarm is accurately generated in the event of an abnormality, a great deal of experience is required of the supervisor. Moreover, considerable manpower is required to set a threshold value for each of a large number of monitoring target areas.
[0006]
FIG. 2 shows the relationship between observation points and discriminant functions in an arbitrary time zone in the conventional determination method. In this method, a monitor sets empirically a threshold value that is formed by a straight line passing through the origin of the BID-ANS characteristics, and when each observation point falls below the threshold value in a certain time zone, it is determined that the communication situation is deteriorated. To do. Thus, since the setting of the threshold depends on the experience of the observer, it is very difficult to automate this threshold based on statistical calculations. Also, since the ABR obtained from the ratio of ANS and BID is non-linear, it is unnatural to compare this with a linear threshold. Furthermore, since the distance (X, Y in the figure) between the observation point and the threshold line varies depending on the BID, the determination accuracy is not constant. That is, since an observation point with a small BID is close to the threshold value, the determination is severe, and conversely, an observation point with a large BID is far from the threshold value, and is a poor determination.
[0007]
As a method for solving such a problem, the following determination method exists.
[0008]
Traffic destined for external carriers is further subdivided for each call type. C 1 the type of call, C 2, ···, C i , C j, ···, C n ( however, C i ∩C j = 0) , a connection request count for each type of call duration T Bid 1, Bid 2, ···, Bid i, ···, Bid n, Ans successful connection count for each type of call Similarly period T 1, Ans 2, ···, Ans i, ··· , Ans n . However, it is assumed that the following assumptions hold regardless of the type of call.
[0009]
At this time, the distribution of e i representing the residual can be approximated by a normal distribution, and the probability density function E i (x) of each e i is obtained from the average value and the variance of the sample data. Bid observations of Bid i i o, ans the observed value of Ans i i o, and e i o = (a i √bid i o + b i) -√ans i o. The probability that the residuals obtained for each type of call are smaller than e 1 o , e 2 o ,... Is as follows.
[Expression 7]
When P (e 1 , e 2 ,...) Falls below a value η determined for each monitored external communication carrier, it is determined that there is a possibility of network abnormality.
[0011]
[Problems to be solved by the invention]
However, such a conventional determination method does not have sufficient determination accuracy, and since it is a determination for each time, there is a possibility that the determination cannot be made when a slightly poor communication situation continues for a long time.
[0012]
Accordingly, an object of the present invention is to provide a communication status determination apparatus in a communication network that has a high determination accuracy and can determine a state in which a somewhat poor communication status continues for a long time.
[0013]
[Means for Solving the Problems]
According to the present invention, the connection success rate (ABR) is obtained from the number of call connection requests (BID) per unit time in the communication network and the number of call connection completions (ANS) per unit time, and the communication status in the communication network. A set of BID and ANS when the communication network is normal is set as sample data for creating a model by binomial distribution , and the average when the communication network is normal from the sample data Based on the model generation means using the binomial distribution for calculating the connection rate, the BID and ANS time series, and the average connection rate generated by the model generation means, the normal hypothesis that the communication network is considered normal One-point likelihood calculation means for calculating one-point likelihood and one-point likelihood of an abnormal hypothesis that the communication network is abnormal, and one-point likelihood of the normal hypothesis and abnormal hypothesis calculated by the one-point likelihood calculation means Series There are communication status determination device in the communication network and a determination means for performing normal or abnormal determination of communication status of a communication network is provided.
[0014]
In the present invention, the relationship between BID and ANS is modeled with a binomial distribution, and the communication status is determined using, for example, the SPRT (Sequential Probability Ratio Test) method. By adopting a configuration for performing such determination processing, automation can be realized and the accuracy of determination is improved. In addition, it is possible to determine a state in which a slightly poor communication state continues for a long time. Further, it is possible to generate a communication determination model that is uniquely determined by flexibly corresponding to the situation of the monitoring target area.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 3 is a block diagram schematically showing the configuration of an embodiment of the communication status judging apparatus of the present invention. In the figure, 11 is a model generation unit, 12 is a one-point likelihood calculation unit, and 13 is an SPRT determination unit. In this embodiment, the relationship between BID and ANS is modeled by a binomial distribution, the communication status is determined using the SPRT method, and a model is generated for each optimal day of the week using the AIC standard.
[0016]
A set S = {(b 1 , a 1 ), (b 2 , a 2 )} of BID and ANS when the communication network is normal is input to the model generation unit 11 as sample data for model creation Is done. At this time, the average connection rate when the communication network is normal
[Equation 9]
The one-point likelihood calculation unit 12 calculates a one-point likelihood for a normal hypothesis that the network is normal and an abnormal hypothesis that is abnormal, for each time. If the observed values of BID and ANS are b and a and the connection success probability is p, the one-point likelihood ll (b, a | p) is as follows.
[Expression 10]
[0018]
The SPRT determination unit 13 determines normality or abnormality based on the SPRT method. This SPRT method is a method of determining normality or abnormality using a logarithmic likelihood time series of a normal hypothesis and an abnormal hypothesis, and the probability of missing an abnormality and the probability of determining normal as abnormal are less than a certain value. As a result, it has the property of minimizing the observation time until a decision is obtained. Therefore, it is suitable for monitoring the communication status.
[0019]
For a binomial distribution, the variance is proportional to the square root of the number of trials. For this reason, it is also conceivable that a regression model is constructed after square root transformation of BID and ANS, and the residual between the regression line and the actual value is approximated by a normal distribution.
[0020]
Table 1 below uses data for 78 days for all external carriers (carriers), and compares the model using the square root transformation as a conventional example with the model using the binomial distribution of the present invention. Yes. However, this comparison is performed using log likelihood.
[0021]
[Table 1]
As can be seen from this table, it can be seen that the logarithmic likelihood of the present invention is greater in all regions and is superior as a model. The one-point likelihood ll (b, a | p w ) at the time of abnormality is a single-point likelihood at the time of normality
[0023]
Next, generation of a day-by-day group model using AIC will be described.
[0024]
Since the connection success rate differs between weekdays and holidays, it is possible to improve the accuracy of the model by creating a model for each day of the week. However, there is a risk of over-learning if too many models are created. It is also necessary to consider that the characteristics of the day of the week differ for each external telecommunications carrier. Therefore, a model was created for each combination of various days of the week, and an optimal day-division method was determined for each external telecommunications carrier using AIC standards.
[0025]
The optimal group structure is shown below.
{Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday, Celebration}
{Monday, Tuesday, Wednesday, Thursday, Friday} {Saturday, Sunday, Celebration}
{Mon, Tue, Wed, Thu, Fri} {Sat, Sun} {Holiday}
{Mon, Tue, Wed, Thu} {Fri} {Sat, Sun, Celebration}
{Month} {Tue, Wed, Thu, Fri} {Sat, Sun, Celebration}
{Month} {Tue, Wed, Thu, Fri} {Sat, Sun} {Holiday}
{Month} {Tue, Wed, Thu, Fri} {Sat} {Sun, Celebration}
{Month} {Tue, Wed, Thu} {Fri} {Sat, Sun, Celebration}
{Month} {Tue, Wed, Thu} {Fri} {Sat} {Sun, Celebration}
{Month} {Tue, Wed, Thu} {Fri} {Sat, Sun} {Holiday}
{Month} {Tue, Wed, Thu} {Fri} {Sat} {Sun} {Holiday}
[0026]
As described above, grouping using the AIC standard is effective because the day of the week reflecting the time difference and culture (for example, Friday is a holiday in the religion) can be divided.
[0027]
In addition, the time at which an abnormality was detected was obtained by changing the value of the sensitivity parameter ε with respect to the BID and ANS time series for all external carriers for the past 78 days.
[0028]
FIG. 5A shows the number of times that a situation that is considered abnormal is detected for various ε. Possible values of ε are numbers greater than 0 and less than 1, and the figure covers the entire range of ε. In the figure, the number of abnormality detections monotonously increases as ε increases. In the case of the conventional method, since the threshold value must be set for each external carrier and for each day of the week, the number of times of abnormality detection cannot be controlled with one parameter. On the other hand, in the present invention, if the expected number of abnormality detections is determined, the value of ε is uniquely determined, and the system operation is facilitated.
[0029]
FIG. 5B shows the relationship between the number of calls to be rescued and ε when the number of calls that should be relieved from the difference between the expected number of connection completions (= BID × p) and the ANS at the time of being regarded as abnormal. Is shown. The figure covers the full range of ε. In the figure, the number of calls to be rescued monotonically increases as ε increases. In the case of the conventional system, the number of rescue call target calls cannot be controlled with one parameter. On the other hand, in the present invention, if the expected number of calls to be repaired is determined, the value of ε is uniquely determined, which facilitates system operation.
[0030]
Furthermore, for the same period, a comparison was made with the case where abnormality detection was performed with a conventional system. For conventional systems, the abnormality detection Demawa number of about 50,000, relieved number of calls during the abnormality detection was about 490,000. On the other hand, in the present invention, when ε = 0.6 is set, the number of abnormality detections is equal to that in the conventional system. However, the number of calls to be repaired in this method is about 750,000, which is the number of calls to be repaired compared to the conventional system. There are about 260,000. Since the traffic abnormality detection of network control for communication relief target call is triggered performed, towards the relieved call number is often present invention at the time of abnormality detection is detection count even approximately equal is the prior art There is a detection effect.
[0031]
All the embodiments described above are illustrative of the present invention and are not intended to be limiting, and the present invention can be implemented in other various modifications and changes. Therefore, the scope of the present invention is defined only by the claims and their equivalents.
[0032]
【The invention's effect】
As described above, according to the present invention, a set of BID and ANS when the communication network is normal is used as sample data for creating a model by binomial distribution , and the average connection when the communication network is normal from the sample data. One point of normal hypothesis that the communication network is considered normal based on the model generation means using the binomial distribution for calculating the rate, the time series of BID and ANS, and the average connection rate generated by the model generation means One-point likelihood calculation means for calculating likelihood and one-point likelihood of abnormal hypothesis that the communication network is considered to be abnormal, and time series of one-point likelihood of normal hypothesis and abnormality hypothesis calculated by one-point likelihood calculation means Is provided with determination means for determining whether the communication status of the communication network is normal or abnormal, so that automation can be realized and the accuracy of the determination is improved. In addition, it is possible to determine a state in which a slightly poor communication state continues for a long time. Further, it is possible to generate a communication determination model that is uniquely determined by flexibly corresponding to the situation of the monitoring target area.
[Brief description of the drawings]
FIG. 1 is a characteristic diagram showing time series data of ABR, BID, and ANS in a daily connection call to a certain country.
FIG. 2 is a diagram showing a relationship between observation points and discriminant functions in an arbitrary time zone in a conventional determination method.
FIG. 3 is a block diagram schematically showing a configuration of an embodiment of a communication status determining apparatus in a communication network according to the present invention.
FIG. 4 is a diagram showing the relationship between observation points and discriminant functions in an arbitrary time zone according to the present invention.
FIG. 5 is a characteristic diagram showing the number of abnormal detections for the sensitivity parameter ε and the number of calls to be relieved in the present invention.
[Explanation of symbols]
11 Model generation unit 12 One-point likelihood calculation unit 13 SPRT determination unit
Claims (5)
通信網が正常時の単位時間当りの呼の接続要求数と単位時間当りの呼の接続完了数との組の集合を２項分布によるモデル作成のためのサンプルデータとし、該サンプルデータから通信網が正常時の平均接続率を算出する２項分布を用いたモデル生成手段と、
呼の接続要求数及び接続完了数の時系列と、前記モデル生成手段によって生成された平均接続率とに基づいて、通信網が正常であると考える正常仮説の一点尤度と、通信網が異常であると考える異常仮説の一点尤度とを算出する一点尤度計算手段と、
該一点尤度計算手段により算出された正常仮説及び異常仮説の一点尤度の時系列を用いて通信網の疎通状況の正常又は異常の判定を行う判定手段とを備えたことを特徴とする通信網における疎通状況判定装置。An apparatus for determining normal or abnormal communication status in a communication network by obtaining a connection success rate from the number of call connection requests per unit time in a communication network and the number of call connections completed per unit time,
A set of sets of the number of call connection requests per unit time and the number of call connections completed per unit time when the communication network is normal is used as sample data for creating a model based on the binomial distribution. A model generation means using a binomial distribution for calculating the average connection rate when
Based on the time series of the number of call connection requests and the number of connection completions and the average connection rate generated by the model generation means, the one-point likelihood of the normal hypothesis that the communication network is normal and the communication network is abnormal A one-point likelihood calculating means for calculating a one-point likelihood of the abnormal hypothesis considered to be
Communication comprising: a determination means for determining normality or abnormality of the communication status of the communication network using a time series of the one-point likelihood of the normal hypothesis and the abnormal hypothesis calculated by the one-point likelihood calculation means Communication status determination device in the network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP04853998A JP3692765B2 (en) | 1997-03-04 | 1998-02-16 | Communication status judging device in communication network |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP9-63965 | 1997-03-04 | ||
| JP6396597 | 1997-03-04 | ||
| JP04853998A JP3692765B2 (en) | 1997-03-04 | 1998-02-16 | Communication status judging device in communication network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPH10308824A JPH10308824A (en) | 1998-11-17 |
| JP3692765B2 true JP3692765B2 (en) | 2005-09-07 |
Family
ID=26388833
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP04853998A Expired - Fee Related JP3692765B2 (en) | 1997-03-04 | 1998-02-16 | Communication status judging device in communication network |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP3692765B2 (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4720295B2 (en) * | 2005-06-02 | 2011-07-13 | 日本電気株式会社 | Abnormality detection system and maintenance system |
| JP5485740B2 (en) * | 2010-02-12 | 2014-05-07 | 株式会社Ｎｔｔドコモ | Failure detection device |
-
1998
- 1998-02-16 JP JP04853998A patent/JP3692765B2/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| JPH10308824A (en) | 1998-11-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6947988B1 (en) | Method and apparatus for allocating resources of a contact center | |
| US8648706B2 (en) | Alarm management system having an escalation strategy | |
| US20190253370A1 (en) | Predicting and updating availability status of a user | |
| US9642562B2 (en) | Bio-surveillance system | |
| US20130060587A1 (en) | Determining best time to reach customers in a multi-channel world ensuring right party contact and increasing interaction likelihood | |
| EP0651549A1 (en) | Telemarketing complex with automatic threshold levels | |
| US7499844B2 (en) | Method and system for predicting network usage in a network having re-occurring usage variations | |
| US9736303B1 (en) | Best time to call parties having multiple contacts | |
| US8750462B2 (en) | Systems and methods for monitoring voice service feature failures in a telecommunication network | |
| CN110290281A (en) | Calling-out method, system, electronic equipment and medium | |
| US9894204B1 (en) | Pacing outbound communications that generate inbound communications | |
| CN108718303B (en) | Safe operation and maintenance management method and system | |
| US6621901B1 (en) | Optimal dynamic agent state assignment | |
| WO2018042239A1 (en) | Techniques for benchmarking pairing strategies in a contact center system | |
| CN110933698A (en) | Monitoring management method, device and equipment for Internet of things card | |
| CA1204488A (en) | Method and apparatus for identifying faulty communications circuits | |
| US20060269050A1 (en) | Adaptive fraud management systems and methods for telecommunications | |
| CA2228591C (en) | System and method for detecting mass addressing events | |
| JP3692765B2 (en) | Communication status judging device in communication network | |
| EP0863649B1 (en) | Apparatus for monitoring state of traffic in telecommunication network | |
| EP3599739A1 (en) | Electronic conferencing system and method of operating an electronic conferencing system | |
| US20100149981A1 (en) | Determining Normal Call Blocking Volume From Call Blast Affecting Trunk Groups | |
| US20060282530A1 (en) | Methods and apparatus for end-user based service monitoring | |
| US20060031461A1 (en) | Deferred method of sending information | |
| CN109508356B (en) | Data abnormality early warning method, device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20040921 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20041005 |
|
| A521 | Written amendment |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20041126 |
|
| 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: 20050531 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20050613 |
|
| R150 | Certificate of patent or registration of utility model |
Free format text: JAPANESE INTERMEDIATE CODE: R150 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20110701 Year of fee payment: 6 |
|
| LAPS | Cancellation because of no payment of annual fees |