JP6293598B2 - Communication traffic prediction apparatus and program - Google Patents

Description

  The present invention relates to a communication traffic prediction apparatus and program, and more particularly to a communication traffic prediction apparatus and program for predicting a future communication traffic amount based on the number of calls of a communication terminal.

  Patent Documents 1-4 disclose techniques for calculating future communication traffic based on actual values of traffic characteristics.

  In Patent Document 1, as a method for predicting the traffic volume of a communication area that fluctuates depending on the date and time, location registration information in a mobile / mobile communication network is monitored and measured, and the call status of a mobile terminal that has made a call is also monitored. A technique for measuring and predicting the future date and time and the call situation for each region based on these past performance values is disclosed.

  In Patent Document 2, as a method for predicting traffic in a fluctuating communication region, traffic depending on each position (unit region) on a map is calculated from traffic in a telecommunication facility and a service providing region of the telecommunication facility. A technique for predicting future position-dependent traffic from actual values of position-dependent traffic is disclosed. The amount of traffic in each unit area is calculated based on the area ratio between the service providing area of the telecommunication equipment and the observation unit area as the traffic between the telecommunication equipment.

  In Patent Document 3, as a method for predicting the communication traffic volume of a fixed network such as a backbone line or an INET line, the time and the size of a packet that flow on the network are measured, and the traffic volume and its fluctuation are affected A technique for managing external factors (calendar information, event status, etc.) and predicting the future traffic volume by applying a statistical analysis method to such information is disclosed.

  In Patent Document 4, as a method for predicting the amount of communication traffic generated when a new service spreads, the number of service subscribers affecting the traffic volume of the new service, the trend / results of the traffic amount at the time of the same service spread in the past, etc. A technique for managing / measuring traffic and predicting the traffic volume at the time of the spread of new services in the future is disclosed.

JP 2004-80315 A JP 2001-168985 A JP 2012-253445 JP JP 2012-182677 A JP 2013-236878 A

  Patent Document 1 predicts the number of calls by measuring signal information related to location registration and signaling, and predicts the amount of communication traffic based on the prediction result of the number of calls. Therefore, if the terminal user repeatedly moves in the boundary region of the location area related to location registration (region / area unit managed by location registration), location registration information is generated more than necessary, and the number of location registration information becomes the number of users. It is not proportional. As a result, the correlation between the number of location registration information and the traffic amount is also lowered, and the prediction accuracy of the traffic amount is lowered.

  Patent Documents 2-4 directly calculate the predicted traffic volume by applying a time-series prediction method to the actual value of the traffic volume. However, the accuracy of time-series prediction decreases as the total amount of communication traffic and the fluctuation range of the traffic amount per communication increase, so traffic that includes a large amount of traffic per communication and a small amount of communication like today. There was a technical problem that the prediction accuracy decreased under the environment.

  Furthermore, in Patent Document 2, although geographical-dependent traffic is measured for each unit area, the traffic amount of each unit area is the total traffic regardless of whether each area is a residential area, a shopping area, an office area, or the like. Calculated as area ratio to quantity. However, since the area ratio and traffic ratio of each unit region are not necessarily equivalent, the traffic volume cannot be predicted with high accuracy.

  Thus, in the prior art, when predicting the future traffic volume, a time series prediction method based on the past or current traffic volume or the like is used. However, if events that greatly fluctuate device demand, such as sales campaigns in-house or other companies, model changes of existing devices, new device sales start, new service (band class) start, etc. are irregularly implemented, Since the number of calls per unit greatly fluctuates, the conventional time series prediction method has a technical problem that the prediction accuracy is lowered.

  In response to such a technical problem, the inventors of the present invention accurately reflect the fluctuation in terminal demand due to an event in the prediction of the number of calls, and based on the prediction of terminal demand that fluctuates according to the event trigger, A communication traffic prediction apparatus capable of accurately predicting the traffic volume for each unit area has been invented and a patent application has been filed (Patent Document 5). However, since Patent Document 5 complicates apportionment calculations for reflecting terminal demand fluctuations in each unit area, a system that can accurately predict future communication traffic for each unit area using a simpler method has been demanded. .

  SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems of the prior art, and a communication traffic prediction apparatus and program capable of accurately predicting the total amount of future communication traffic based on the number of calls regardless of the presence or absence of various events that fluctuate terminal demand Is to provide.

  In order to achieve the above object, the present invention is characterized in that a communication traffic prediction apparatus for predicting the amount of communication traffic by a communication terminal has the following configuration.

  (1) Traffic characteristic measurement means that measures traffic characteristics including the number of calls for each unit area obtained by dividing the traffic prediction area according to geography, and the contribution ratio of the number of calls in each unit area to the fluctuation in the number of terminals in the entire prediction area Based on traffic characteristics, call number prediction means for predicting the future number of calls for each unit area, future terminal demand for the entire prediction area, actual values and terminal demands fluctuating Terminal demand prediction means for predicting based on the schedule of the event to be performed and the sales season (hereinafter referred to as event), and correction means for correcting the prediction result of the number of calls based on the contribution rate and the prediction result of terminal demand Equipped.

  (2) Predicted traffic total amount calculating means for calculating a predicted traffic total amount based on the corrected call number prediction result for each unit area is provided.

  (3) The terminal demand prediction means further includes terminal attribute reflection means for reflecting the terminal dependence of the number of calls in the prediction result.

  (4) The terminal demand prediction means further includes a geographical attribute reflection means for predicting terminal demand for each unit area by reflecting the geographical dependence of the number of calls in the prediction result.

According to the present invention, the following effects are achieved.
(1) Since fluctuations in terminal demand due to events can be accurately reflected in the prediction of the number of calls for each unit area without using complex apportionment calculations, future communication traffic is based on the prediction of terminal demands that fluctuate according to event triggers. The amount can be accurately predicted for each unit area.

  (2) Since the terminal demand is predicted in consideration of the event attribute, the future terminal demand can be accurately predicted even when the terminal demand fluctuation of the event trigger has a dependency on the event attribute.

  (3) When correcting the predicted number of calls based on the predicted future terminal demand, when converting the predicted number of terminals to the number of calls, the number of calls per unit time or per terminal (number of calls / unit ) Device dependency and geographic dependency are reflected, so even if the number of calls / device has device dependency or geographical dependency, the call number prediction result can be accurately corrected based on the terminal demand prediction result It becomes like this.

It is the functional block diagram which showed the structure of the principal part of the communication traffic prediction apparatus with which this invention is applied. It is the figure which showed typically the operation | movement of the traffic estimation part. It is the figure which showed the relationship between the prediction range by a traffic prediction part and a terminal demand prediction part, and the performance range used for the said prediction. It is the functional block diagram which showed the structure of the terminal demand prediction part. It is the figure which showed an example of the prediction result of a terminal demand. It is the functional block diagram which showed the other structure of the terminal demand prediction part. It is the flowchart which showed the operation | movement of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Here, an outline of the present invention will be described first, and then an embodiment of the present invention will be described in detail.

  The total amount of communication traffic to be predicted by the present invention is determined by the product of the number of calls (attempt number) and the amount of traffic per call. Therefore, if the future number of calls can be accurately predicted, the prediction accuracy of the total amount of traffic can be improved. be able to. Here, in the time series prediction generally used for prediction of time series data such as daily and hourly number of calls, future values are predicted based on past results / trends. If the future value fluctuates below, the prediction accuracy deteriorates.

  The factors that cause the number of calls to fluctuate above or below the past trend are: (1) the number of devices sold has increased or decreased significantly over the past trend, and (2) the number of subscribers has increased significantly over the past trend / (3) The number of devices sold or canceled has been biased toward specific terminals more than the past trend, and (4) The number of terminals compatible with specific communication methods and band classes has increased or decreased more than the past trend. However, (5) new communication methods and band-class compatible terminals may be newly sold that cannot appear in past trends / results.

  These call fluctuation factors are due to the presence or absence of events that lead to an increase or decrease in device demand, such as sales campaigns in-house or other companies, model changes to existing devices, new device sales, and new services (band class). Strongly dependent.

  In other words, since the number of terminals and the number of calls are approximately proportional, if the future terminal demand is predicted in consideration of scheduled events and this can be reflected in the prediction of the number of calls, the impact of each of the above variable factors This makes it possible to predict the number of calls with high accuracy.

  Therefore, in the present invention, the number of calls is calculated by predicting the number of calls in the future by a general prediction method such as time series prediction, while the terminal of the future is based on the contents, timing, scale, etc. of the scheduled event. By predicting demand and correcting the predicted number of calls based on the prediction result of terminal demand, the prediction accuracy of the number of calls is improved, and thus the prediction accuracy of the total communication traffic is improved.

  Note that the proportional relationship between the number of terminals and the number of calls is terminal-dependent or geographical-dependent, and the rate of increase in the number of calls with respect to the rate of increase in the number of terminals is not uniform, and terminal attributes such as geographical attributes, communication methods, and band classes May vary depending on.

  In addition, future terminal demand is also geographically dependent. For example, if attention is focused on smartphones as communication terminals, a large increase in future demand cannot be expected regardless of sales promotion campaigns, etc. In insufficient regions, there may be cases where a large increase in demand can be expected in the future due to promotional campaigns. Or even if the current penetration rate is the same, the future terminal demand is not always the same.

  Therefore, in the present invention, when a future terminal demand is predicted in consideration of a scheduled event and a future number of calls is corrected, a function that converts the number of terminals into the number of calls is converted into a terminal dependency and a geographical dependency. I was able to reflect.

  FIG. 1 is a functional block diagram showing a configuration of a main part of a communication traffic prediction apparatus according to an embodiment of the present invention, which is configured by mounting an application (program) for realizing each function on a general-purpose computer or server. Alternatively, it may be configured as a dedicated machine or a single-function machine in which a part of the application is implemented in hardware or ROM. Here, illustrations of components unnecessary for the description of the present invention are omitted.

  The communication log collection unit 8 with position information was seen as the communication log of each communication terminal when time information, transmission source information, destination information, and each communication terminal communicated for each call (communication trial unit) (communication possible) Various information including communication delay times between the base station and each base station is collected, and further estimated by the positioning result using each communication delay time or the GPS function installed in each communication terminal. The recorded position is recorded as position information.

  The communication log DB1 with position information defines a large number of geographically dependent unit areas by virtually dividing the prediction area of communication traffic into a mesh shape based on separately divided information, and each communication log is Based on the position information, it is linked to one of the unit areas to create a database. The division information includes a prediction target area and information on a mesh standard / unit (size) defined by the Geographical Survey Institute or the Ministry of Internal Affairs and Communications, for example.

  Based on the communication log associated with each unit area, the traffic characteristic measurement unit 2 determines the traffic characteristics including the number of calls Ci (i is a unit area identifier) and the traffic amount Vi / Ci per call for each unit area. taking measurement.

  As schematically shown in FIG. 2, the traffic prediction unit 4 applies an appropriate prediction method to the measurement results up to the present of the number of calls Ci obtained for each unit area (the number of calls in the future). Cfi is predicted for each unit area and unit time.

  As a method for predicting the number of calls, for example, a time series prediction method such as Autoregressive (autoregressive), Moving Average (moving average), or ARIMA (Autoregressive Integrated Moving Average) can be applied. Alternatively, a time series prediction method such as Seasonal ARIMA may be applied in consideration of Seasonal components (seasonal and periodic fluctuations) and automatically calculating specific parameters from past performance data and reflecting them in the prediction. Furthermore, in addition to the above time series prediction methods, prediction methods such as SVR (support vector regression), NN (neural network), genetic algorithm, and machine learning algorithm may be applied.

  The unit time may be a fixed time (for example, a one-hour cycle), or a unit region with a large amount of traffic is set to a shorter time cycle, while a unit region with a small amount of traffic is set to a longer time cycle. For example, it may be changed dynamically according to the traffic characteristics of each unit area.

  Returning to FIG. 1, time series history information related to the demand P of the communication terminal is registered in the sales record DB 3. The terminal demand prediction unit 5 predicts a future terminal demand (predicted terminal demand) Pf that fluctuates non-stationarily by executing a scheduled event based on the history information of the terminal demand P.

  FIG. 3 is a diagram showing the relationship between the prediction range by each of the prediction units 4 and 5 and the actual range used for the prediction. In the present embodiment, the future number of calls and terminal demand that fluctuate due to the current event (current) are the same as or similar to the current event, but the number of calls ΔCi and the terminal demand Based on the actual value of the number of fluctuations ΔP, it is calculated for each prediction period that is cumulatively extended in a predetermined cycle.

  That is, predicted values Cfi_10 and Pf_10 10 days after the present are calculated based on the numbers of fluctuations ΔP_10 and ΔCi_10 observed in the first period (for example, 10 days) from the past event occurrence date. Furthermore, predicted values Cfi_20 and Pf_20 after 20 days from the present are calculated based on the number of fluctuations ΔP_20 and ΔCi_20 for two periods (20 days in total) in which 10 days are accumulated. Further, predicted values Cfi_30 and Pf_30 after 30 days from the present are calculated based on the number of fluctuations ΔP_30 and ΔCi_30 for three periods (total 30 days) in which 10 days are accumulated.

Returning to FIG. 1, the contribution rate calculation unit 9 calculates the contribution rate of the variation in the number of calls in each unit area in the past variation in the number of terminals in the entire prediction area. In this embodiment, if the number of terminal fluctuations in the entire prediction area is ΔP and the number of call fluctuations in each unit area i is ΔCi, the call in each unit area i in the terminal number fluctuation ΔP in the entire prediction area The contribution ratio Ri of the number variation ΔCi is obtained by the following equation (1).
Ri = ΔCi / ΔP… (1)

  Here, the prediction area is not particularly limited as long as it is a certain wide area that can include a plurality of unit areas i, and it may be a branch office unit that manages sales plans and network facilities, as well as administrative and geographic areas such as Tohoku and Kanto. Basic units may be used.

The correction coefficient calculation unit 10 calculates a correction coefficient Ki for reflecting the contribution rate Ri and the predicted terminal demand Pf to the predicted number of calls Cfi calculated for each unit area i based on the following equation (2). Set for each.
Ki = 1 + (Ri × Pf) (2)

  The predicted call number correction unit 6 corrects the predicted call number Cfi of each unit area i based on the correction coefficient Ki, and outputs this as the corrected predicted call number Cfpi. In the present embodiment, the number of predicted calls Cfi corresponding to each unit area i is multiplied by the correction coefficient Ki.

  The predicted traffic total amount calculation unit 7 multiplies the corrected predicted number of calls Cfpi after each unit time by the statistical value (actual value) Vi / Ci of the traffic amount per call, thereby calculating the future predicted traffic total amount Vfi ( = Cfpi × Vi / Ci) is calculated for each unit area.

  Next, the configuration and function of the terminal demand prediction unit 5 will be described in detail. The terminal demand forecasting unit 5 multiplies the prediction result to correct the prediction result of the number of calls in order to accurately predict the future total amount of communication traffic regardless of irregular events such as sales promotion campaigns and sales of new products. The rate Pf was calculated.

  Therefore, firstly, the fluctuation of terminal demand at the scheduled future event will be the same as when the equivalent event was implemented in the past, and second, the terminal demand observed at the past event It was assumed that the change includes a steady fluctuation component that is unrelated to the existence of the event.

  Then, by applying the actual value of the terminal demand to, for example, the same time series prediction algorithm as the above-mentioned number of calls prediction, while predicting a future steady terminal demand change, the same event at the scheduled event timing By reflecting the actual value of the change in terminal demand when this is implemented in the past, it is possible to accurately predict future terminal demand that fluctuates unsteadily due to event triggers.

  Based on the assumption that the fluctuation in the number of calls is proportional to the fluctuation in the number of terminals, the future terminal demand predicted as the non-stationary fluctuation of the event trigger is calculated based on the time-series prediction results (steady state The correction value Pf was obtained by dividing by the prediction result based on typical terminal demand change.

  FIG. 4 is a functional block diagram showing the configuration of the terminal demand forecasting unit 5. In this embodiment, terminal demand that fluctuates non-steadily at an event trigger is attributed to the attribute information of a scheduled event. It is characterized in that it is predicted.

  In the terminal demand prediction unit 5, the event schedule registration unit 50 includes the event type, the implementation date, the implementation period, and the implementation scale as the attribute information (event attribute) of the scheduled event by, for example, an operator's manual operation. Entered.

  The first prediction unit 51 acquires time-series data P of the terminal demand history from the sales performance DB 3 and applies the first prediction algorithm to the terminal demand history time series data P to obtain the future steady terminal demand (first terminal demand). Predict. As the first prediction algorithm, a time series prediction method such as Autoregressive (Autoregressive), Moving Average (Moving Average), or ARIMA (Autoregressive Integrated Moving Average) can be used similarly to the algorithm for predicting the number of calls. .

  Based on the input event attributes, the prediction model construction unit 53 relates to the sales performance when a similar event having a similar event type, implementation date age, implementation period, implementation scale, etc. has been executed in the past. It is different from the first prediction algorithm such as a forecast model or popular curve (for example, logistic curve or Gompertz curve, etc.) that predicts terminal demand when scheduled data is selected and acquired from series sales data DB3 A second prediction algorithm (hereinafter represented by a prediction model) is constructed. The second prediction unit 52 predicts future terminal demand (second terminal demand) that fluctuates unsteadily at a scheduled event trigger based on the constructed prediction model.

  FIG. 5 is a diagram illustrating an example of a prediction result by the first and second prediction units 51 and 52, and a change in terminal demand due to an event with respect to the first prediction result (broken line) by the first prediction unit 51. The second prediction result (solid line) by the second prediction unit 52 considering (in this case, increase) is upside down.

  Since this second prediction result also includes a time-series prediction component (first terminal demand), the terminal demand calculation unit 54 divides the second terminal demand by the first terminal demand, for example, to generate the second prediction. The number of calls is calculated by offsetting the time-series forecast component from the result, obtaining the rate of increase / decrease in the future terminal demand with respect to the first terminal demand forecast result, and applying this to a predetermined function for calculating the number of calls from the number of terminals Increase / decrease rate Pf.

  For example, as shown in the following equation (3), the increase / decrease rate Pf of the number of calls can be calculated as a division between the predicted value based on the demand prediction using the event DB and the time series prediction using the actual demand. Alternatively, if it can be calculated as an index that indicates how much the future trend of increase / decrease compared to the trend of increase / decrease in the past, the terminal increase / decrease in the past several months using the actual demand as shown in the following equation (4) The number or ratio may be calculated and calculated as a division of the number of predicted values or the ratio of the predicted value based on demand prediction using the event DB.

    Pf = Demand forecast value / time series forecast value based on event schedule… (3)

    Pf = Increase / decrease number (or ratio) of demand forecast value based on event schedule / Increase / decrease number (or ratio) of time-series actual values for the past few months… (4)

  According to the present embodiment, it is possible to accurately predict future terminal demands that change in response to a scheduled event. In addition, according to the present embodiment, since terminal demand is predicted in consideration of event attributes, it is possible to accurately predict future terminal demand even when terminal demand fluctuations of event triggers have dependency on event attributes. become.

  FIG. 6 is a functional block diagram showing another configuration of the terminal demand prediction unit 5, and the same reference numerals as those described above represent the same or equivalent parts. The present embodiment is characterized in that the prediction model construction unit 53 includes a terminal attribute reflection unit 53a that reflects the terminal dependence of the number of calls in the function for obtaining the increase / decrease rate Pf from the terminal demand prediction result.

  In other words, when attention is paid to communication methods as terminal attributes, the number of calls per terminal tends to be higher in LTE terminals than in 3G terminals. For example, the number of calls per unit time of 3G terminals is about 10 calls. For example, that of an LTE terminal is about 15 calls.

  Therefore, assuming that the terminal demand will increase from 1 million to 1.1 million for 3G terminals and 100,000 to 1.1 million for LTE terminals due to events such as promotional campaigns, the increase in the number of terminals will double ( The number of calls will increase from 10 million calls to 11 million calls on 3G terminals, while it will increase from 1.5 million calls to 16.5 million calls on LTE terminals. This is a 2.4-fold increase to 27.5 million calls.

  As described above, since the proportional relationship between the number of terminals and the number of calls is terminal-dependent, in the present embodiment, the terminal attribute reflecting unit 53a is provided, and the ratio of LTE terminals to the estimated terminal demand is the latest. By predicting from the terminal sales situation, etc., and reviewing the relationship between the terminal demand and the rate of change Pf according to the ratio, the increase in terminal demand can be correctly reflected in the increase in the number of calls.

  According to the present embodiment, even when the proportional relationship between the number of terminals and the number of calls is terminal-dependent, it is possible to accurately predict the number of calls caused by the terminal demand fluctuation at the event trigger.

  FIG. 7 is a flowchart showing the operation of the present embodiment. In step S1, the traffic characteristic measurement unit 2 determines the unit area based on the communication log stored in the communication log DB1 with position information. The actual value of the number of calls Ci is measured, and the statistical value of the traffic volume Vi / Ci per call is calculated.

  In step S2, the contribution rate calculation unit 9 calculates the contribution rate Ri of the call number variation in each unit region in the past variation in the number of terminals in the entire prediction area based on the above equation (1). In step S3, the correction coefficient Ki for reflecting the contribution ratio Ri and the predicted terminal demand Pf to the predicted call number Cfi calculated for each unit area i by the correction coefficient calculation unit 10 is expressed by the above equation (2). Based on the unit area i, it is set.

  In step S4, the traffic prediction unit 3 calculates a future predicted number of calls Cf by applying an appropriate prediction method to the actual value of the number of calls C. In step S5, the terminal demand prediction unit 5 calculates, for example, a terminal demand predicted increase rate as the terminal demand Pf based on the history information regarding the sales performance P of the communication terminal accumulated in the sales performance DB3. In step S6, the predicted call number Cfi is corrected based on the correction coefficient Ki to obtain a corrected predicted call number Cfpi.

  In step S7, the predicted traffic total amount calculation unit 4 calculates the predicted traffic total amount Vfi by multiplying the corrected predicted call count Cfpi for each unit area by the traffic amount statistical value Vi / Ci per call. At this time, if the number of calls Ci is a predicted value in units of one hour, the predicted traffic total amount Vfi is also calculated in the unit area and every hour. For the statistical value Vi / Ci, an average value, median value, predetermined percentage value (% ile, for example, 99% ile) or the like may be used to remove abnormal values and noise components.

  DESCRIPTION OF SYMBOLS 1 ... Communication log DB with position information, 2 ... Traffic characteristic measurement part, 3 ... Sales performance DB, 4 ... Traffic prediction part, 5 ... Terminal demand prediction part, 6 ... Predicted call number correction part, 7 ... Predicted traffic total amount calculation part , 8 ... Communication log collection unit with location information, 9 ... Contribution rate calculation unit, 10 ... Correction coefficient calculation unit, 50 ... Event schedule registration unit, 51 ... First prediction unit, 52 ... Second prediction unit, 53 ... Prediction model Construction unit, 53a ... Terminal attribute reflection unit, 53b ... Geographic attribute reflection unit, 54 ... Terminal demand calculation unit

Claims (11)

In a communication traffic prediction device that predicts the amount of communication traffic by a communication terminal,
Traffic characteristic measuring means for measuring traffic characteristics including the number of calls for each unit area obtained by dividing the traffic prediction area according to geography,
Contribution rate calculating means for calculating the contribution rate of the call number variation of each unit area in the terminal number variation of the entire prediction area;
Call number predicting means for predicting the future number of calls for each unit area based on the traffic characteristics;
Terminal demand prediction means for predicting the future terminal demand of the entire prediction area based on the actual value and the schedule of the event that fluctuates the terminal demand;
A communication traffic prediction apparatus comprising correction means for correcting the prediction result of the number of calls based on the contribution rate and the prediction result of terminal demand.   The communication traffic prediction apparatus according to claim 1, further comprising a predicted traffic total amount calculation unit that calculates a predicted traffic total amount based on the corrected number of calls predicted result for each unit area. Further comprising event schedule registration means for registering the attribute of the scheduled event;
The terminal demand forecasting means is
A first prediction means for predicting the first terminal demand by applying the actual value of the terminal demand to the first prediction algorithm;
The second terminal demand triggered by the event based on the terminal demand observed when the event having the same attribute as the scheduled event has been executed in the past is different from the first prediction algorithm. Second prediction means for predicting with a prediction algorithm of 2,
The communication traffic prediction apparatus according to claim 1, further comprising terminal demand calculation means for calculating a future terminal demand based on the first and second terminal demands.   The communication traffic prediction apparatus according to claim 3, wherein the first prediction algorithm is the same as or equivalent to the prediction algorithm of the call number prediction means.   The communication traffic prediction apparatus according to claim 3 or 4, wherein the event attributes include at least one of an event type, an implementation date, an implementation period, and an implementation scale.   6. The communication traffic prediction apparatus according to claim 3, wherein the second prediction means further comprises terminal attribute reflection means for reflecting the terminal dependence of the number of calls in the prediction result.   The said 2nd prediction means was further provided with the geographical attribute reflection means which reflects the geographical dependence of the number of calls in a prediction result, and estimates a terminal demand for every unit area | region. The communication traffic prediction apparatus according to claim 1. The communication traffic prediction apparatus according to claim 1, wherein the unit area is a mesh area obtained by dividing the prediction area into a mesh shape.   4. The communication traffic prediction apparatus according to claim 3, wherein the terminal demand calculation unit divides the second terminal demand by the first terminal demand. The second terminal demand is a predicted value based on the event schedule of the communication terminal, or an increase / decrease number of the predicted value or an increase / decrease ratio,
The communication traffic prediction apparatus according to claim 9, wherein the first terminal demand is a time-series actual value of a communication terminal or an increase / decrease number or an increase / decrease ratio of the time-series actual value. In a communication traffic prediction program for predicting the amount of communication traffic by a communication terminal,
Procedures for measuring traffic characteristics including the number of calls for each unit area obtained by dividing the traffic prediction area according to geography,
A procedure for calculating the contribution rate of the number of calls in each unit area in the number of terminals in the entire prediction area;
A procedure for predicting the future number of calls for each unit area based on the traffic characteristics;
A procedure for predicting the future terminal demand of the entire prediction area based on the actual value and the schedule of the event that fluctuates the terminal demand;
A communication traffic prediction program that causes a computer to execute a procedure of correcting the prediction result of the number of calls based on the contribution rate and the prediction result of terminal demand.