JP6560644B2 - Waiting time evaluation apparatus and waiting time evaluation method - Google Patents

Description

  The present invention relates to QoE evaluation for data communication.

  Currently, the web is widely used in various fields. Web user experience (Quality of Experience: QoE) is measured and evaluated based on indicators to show comfort of use such as web page display speed. In Non-Patent Document 1, web QoE is determined based on a reference value or threshold value of a uniform display speed regardless of web page contents. Non-Patent Document 2 proposes a technique for collecting the waiting time of each web page.

The Cost of Poor Web Performance [INFOGRAPHIC], November 30, 2012 (http://blog.smartbear.com/web-performance/the-cost-of-poor-web-performance-infographic/) Navigation Timing, W3C Recommendation 17 December 2012 (http://www.w3.org/TR/navigation-timing/)

  Since the web is composed of various contents, the time required for display (hereinafter referred to as “display waiting time”) may vary depending on the page even in the same environment. The state that QoE wants to judge is deteriorated because the content is originally not heavy and delayed, but is slower than usual due to network delay, server congestion, increase in processing load on the terminal, and the like.

  However, according to the prior art, this determination is considered difficult. In other words, as with the prior art, if the degradation criterion is based on a display latency threshold that is uniformly set without considering the difference in web page content, the content is heavy and degradation due to network and server load has occurred. There is a problem that it is judged that the web page is deteriorated even though it is not present, or it is not judged that it is degraded as long as it is within the threshold even if the light web page is delayed significantly.

  For example, if the waiting time measurement value population of each page is set to be deteriorated in the upper 5% of waiting time, etc., the deterioration may be missed due to the quality of the environment when collecting observation samples, On the other hand, there is a possibility that a non-deteriorated one is determined to be deteriorated.

  Further, although a deterioration criterion of X times the standard deviation from the average value of the waiting time is possible, it is difficult to determine the threshold value X indicating where the deterioration is. In addition, when a long waiting time is included in the observation sample, it is not appropriate to evaluate with an average value.

  In view of the above-described problems, an object of the present invention is to provide a waiting time evaluation apparatus and method for appropriately evaluating the waiting time of a web page.

  In order to solve the above problems, one aspect of the present invention is a latency evaluation device that evaluates the latency of a web page, the data acquisition unit that acquires a measurement value group of display latency of a target web page, The present invention relates to a waiting time evaluation apparatus including a deterioration determination unit that determines that the waiting time of the target web page is deteriorated when a measured value portion according to a power distribution is detected in the acquired measured value group of display waiting time.

  ADVANTAGE OF THE INVENTION According to this invention, the waiting time evaluation apparatus and method which evaluate appropriately the waiting time of a web page can be provided.

FIG. 1 is a block diagram showing a functional configuration of a waiting time evaluation apparatus according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing a power line fitting process for a cumulative complementary distribution according to an embodiment of the present invention. FIG. 3 is a flowchart showing the waiting time evaluation process according to one embodiment of the present invention. FIG. 4 is a block diagram showing a hardware configuration of a waiting time evaluation apparatus according to an embodiment of the present invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  In the following embodiments, a waiting time evaluation apparatus and method for evaluating the waiting time of a web page based on the display waiting time of the web page are disclosed. The following example can be summarized as follows: When the web page waiting time deteriorates, such as when the web page display waiting time increases, the degraded part and the part that has not deteriorated have different statistical distribution structures. Is done. Based on these assumptions, it is possible to determine whether each reading attempt of the target web page has deteriorated without depending on the content amount of each web page, the network status at the time of evaluation sample collection, etc. Become.

  Specifically, it is considered that the display waiting time follows a power distribution in a deteriorated portion, while the display waiting time follows a different statistical distribution such as a normal distribution in an undegraded portion. Based on such consideration, it is determined whether or not the waiting time of the web page is deteriorated by specifying whether or not there is a measured value portion that follows the power distribution in the display waiting time measurement value group of the target web page. In the following embodiments, the logarithmic transformation can be used to logarithmically convert the measurement value group using the property that the power function can be expressed by a straight line (power straight line), and the converted measurement value can be applied to the straight line. Determine if there is.

  First, with reference to FIGS. 1-2, the waiting time evaluation apparatus by one Example of this invention is demonstrated. FIG. 1 is a block diagram showing a functional configuration of a waiting time evaluation apparatus according to an embodiment of the present invention.

  As illustrated in FIG. 1, the waiting time evaluation apparatus 100 includes a data acquisition unit 110 and a deterioration determination unit 120.

  The data acquisition unit 110 acquires a measurement value group of the display waiting time of the target web page. Specifically, the data acquisition unit 110 receives, as input data, a sample group of display latency measurement results required to display a web page to be evaluated on a terminal or the like, or sets the display latency by itself. You may acquire the sample group of a measurement result by measuring. Thereafter, the data acquisition unit 110 sorts the acquired measurement value group in ascending order for the subsequent processing. For example, when N measurement values are acquired, the data acquisition unit 110 configures a waiting time measurement value list T = [t_1, t_2, ..., t_N] by sorting the acquired measurement value group in ascending order, Provided to the deterioration determination unit 120.

  In addition, the data acquisition unit 110 receives, as input parameters, various parameters used for the straight line fitting process for the measurement value group described in detail below. The parameters include the slope R_p of the power line, the fitting allowable width γ, the fitting cut-off width β, the endpoint cut-off threshold X_th, the determination lower-limit threshold μ, the determination upper-limit threshold κ, and the like as described later. Note that the slope R_p of the power function may be provided as an input parameter, as described later, or may be calculated by performing a least square regression on the measurement value group.

  The degradation determination unit 120 determines that the waiting time of the target web page has deteriorated when detecting a measured value portion that follows the power distribution in the acquired measured value group of display waiting time. In the present embodiment, the degradation determination unit 120 determines whether the logarithmically expressed display wait time measurement value group and its cumulative complementary distribution can be applied to a straight line representing a power function (power line). Specifically, the deterioration determination unit 120 performs logarithmic conversion on the measurement value portion extracted from the acquired measurement value group of the display wait time, whereby the wait time and the cumulative complementary distribution as illustrated in FIG. The cumulative complementary distribution data consisting of a data group indicating the relationship between the generated cumulative complementary distribution data is generated, and it is determined whether the generated cumulative complementary distribution data can be applied to the power line. Judged to follow the power distribution.

  A specific fitting process according to an embodiment will be described.

  First, as described above, the data acquisition unit 110 constructs a list of waiting time measurement values T = [t_1, t_2,... The fitting cut-off width β, the endpoint cut-off threshold value X_th, the determination lower-limit threshold value μ, and the determination upper-limit threshold value κ are received as input parameters.

  Here, as shown in FIG. 2, the slope R_p of the power line is negative (R_p <0), and may be −2.2, for example.

  The fitting allowable width γ represents a fitting allowable range in which a measured value within the allowable width is regarded as following a power line, and may be a positive number such as 0.2, for example.

  The fitting cut-off width β represents the ratio (0 ≦ β <1) of the number of allowable values that are allowed to be out of the fitting allowable range in the measurement value group to be fitted. For example, 0.05 (= 5%) There may be. As shown in the drawing, when a measurement value that falls outside the fitting allowable range appears beyond the fitting cut-off width β, the fitting process is terminated, and the measurement value immediately before the appearing measurement value may be set as the fitting range.

  The endpoint truncation threshold value X_th represents a value for truncating an end point portion (0 <X_th <1) that is not suitable for evaluation as an abnormal value among N measurement value groups. For example, when 1% of abnormal values are included If estimated, set to 0.99.

  The determination lower limit threshold μ is a value (0 <μ <1) that determines that the measured value group does not apply to the power line and the waiting time has not deteriorated if the measurement value is applied to the power line only at or below the threshold. For example, 0.03.

  The determination upper limit threshold κ represents an upper limit (0 <κ <1) of a deterioration threshold described later, and may be 0.5, for example. If the measured value falls within the power line above the threshold, the degradation threshold cannot be determined because the distribution range of the power distribution is too large and is not appropriate for determining whether the degradation is originally intended. Means that.

When the waiting time measurement value list T and the input parameter are received, the deterioration determination unit 120 first calculates an abnormal value of the distribution tail from the waiting time measurement value list T = [t_1, t_2, ..., t_N] according to the end point truncation threshold value X_th. truncate. Specifically, the deterioration determination unit 120
Nr = int ((N-1) * X_th) +1
And Nr evaluation target lists T ′ = [t_1, t_2,..., T_Nr] are constructed. Here, int () is a function that truncates the decimal part of the argument of the real value.

Next, the deterioration determination unit 120 generates cumulative complementary distribution data as shown in FIG. 2 for the list T ′. Specifically, the deterioration determination unit 120
Xc = map (log, T ')
And the logarithmic value of each element of the list T ′ is calculated. Here, map () is a function that applies a logarithmic function log to each element of T ′. Furthermore, the deterioration determination unit 120
Tc = map (cologN, Tidx ')
Calculate Here, Tidx ′ represents the subscript string [1, 2,..., Nr] of T ′, and cologN calculates the cumulative complementary distribution of each element i (i = 1, 2,..., Nr) of T ′. Is a function for
cof (i) = log (1- (i-1.0) / (N-1.0))
Is applied to each element i (i = 1, 2, ..., Nr) of Tidx '.

Next, the deterioration determination unit 120 specifies a measurement value that falls outside the fitting allowable range by the fitting cut-off width β or more in the search range based on the determination lower limit threshold μ and the determination upper limit threshold κ, and the measurement value immediately before the specified measurement value ( (Deterioration threshold). Specifically, the deterioration determination unit 120
s = int (μ * N), k = int (κ * N)
To calculate the lower limit s and the upper limit k of the search range.

  In addition, the deterioration determination unit 120 is a measured value portion of the display waiting time that falls within the allowable range of the straight line defined by the slope R_p and the fitting allowable width γ from the rear of the column of Xc and Tc, that is, from the long waiting time. Explore. Specifically, the deterioration determination unit 120 generates a vector X2 (length is Nr) obtained by flattening Xc with a slope R_p. That is, the value of X2 for the i-th measurement value is as follows.

X2 [i] = Xc [i] -R_p * Tc [i]
Thereafter, the degradation determination unit 120 sequentially searches for an element of Xc having the minimum waiting time that fits the power line from the rear of the column of Xc according to the following procedure.

ここで、Xmin, Xmaxはそれぞれ、X2'の分布の裾をβ/2ずつ切り捨てた残りの最小値及び最大値であり、例えば、β=0.05 (=5%)である場合、Xmin, Xmaxはそれぞれ２．５パーセンタイル及び９７．５パーセンタイルとなる。また、if条件が成立した場合、(i-1)番目までは許容範囲内にあり、劣化判定部１２０は、劣化している範囲を示す劣化閾値として(i-1)番目の値を出力する。他方、if条件が成立しない場合、劣化判定部１２０は、カウンタiをインクリメントし、forループを繰り返す。

Here, Xmin and Xmax are the remaining minimum and maximum values obtained by rounding down the tail of the distribution of X2 ′ by β / 2, for example, when β = 0.05 (= 5%), Xmin and Xmax are The 2.5 percentile and 97.5 percentile respectively. Also, if the if condition is satisfied, the (i-1) th is within the allowable range, and the deterioration determining unit 120 outputs the (i-1) th value as the deterioration threshold indicating the deteriorated range. . On the other hand, when the if condition is not satisfied, the deterioration determination unit 120 increments the counter i and repeats the for loop.

  The deterioration determination unit 120 determines that a measurement value portion that matches a power line is detected when D is a value obtained by the above-described procedure and K> D and D> s, and the power value distribution is determined as a power distribution. It is determined that the waiting time of the target web page has deteriorated. Furthermore, the deterioration determination unit 120 outputs a deterioration threshold value T [Nr-D], and the measurement value portion is the (Nr-D) th and subsequent measurement values in the list T, and the deterioration threshold value T [Nr-D]. It can be determined that the above measured values are deteriorated.

  On the other hand, when D ≦ s, the deterioration determination unit 120 determines that there is no measurement value portion that applies to the power line in the measurement value group, and that the waiting time has not deteriorated.

  When k ≦ D, the deterioration determination unit 120 determines that the waiting time deterioration has occurred for a large number of measurement values. In this case, the deterioration determination unit 120 does not output the deterioration threshold because it cannot specify the deterioration threshold.

  A specific fitting process according to another embodiment will be described. This embodiment can be applied to a case where the data acquisition unit 110 is not provided with the slope R_p of the power line in advance.

  First, as described above, the data acquisition unit 110 constructs a list of waiting time measurement values T = [t_1, t_2,. β, end point truncation threshold value X_th, determination lower limit threshold value μ, and determination upper limit threshold value κ are received as input parameters.

When the waiting time measurement value list T and the input parameter are received, the deterioration determination unit 120 first calculates an abnormal value of the distribution tail from the waiting time measurement value list T = [t_1, t_2, ..., t_N] according to the end point truncation threshold value X_th. truncate. Specifically, the deterioration determination unit 120
Nr = int ((N-1) * X_th) +1
And Nr evaluation target lists T ′ = [t_1, t_2,..., T_Nr] are constructed. Here, int () is a function that truncates the decimal part of the argument of the real value.

Next, the deterioration determination unit 120 generates cumulative complementary distribution data as shown in FIG. 2 for the list T ′. Specifically, the deterioration determination unit 120
Xc = map (log, T ')
And the logarithmic value of each element of the list T ′ is calculated. Here, map () is a function that applies a logarithmic function log to each element of T ′. Furthermore, the deterioration determination unit 120
Tc = map (cologN, Tidx ')
Calculate Here, Tidx ′ represents the subscript string [1, 2,..., Nr] of T ′, and cologN calculates the cumulative complementary distribution of each element i (i = 1, 2,..., Nr) of T ′. Is a function for
cof (i) = log (1- (i-1.0) / (N-1.0))
Is applied to each element i (i = 1, 2, ..., Nr) of Tidx '.

Next, the deterioration determination unit 120 specifies a measurement value that falls outside the fitting allowable range by the fitting cut-off width β or more in the search range based on the determination lower limit threshold μ and the determination upper limit threshold κ, and the measurement value immediately before the specified measurement value ( (Deterioration threshold). Specifically, the deterioration determination unit 120
s = int (μ * N), k = int (κ * N)
To calculate the lower limit s and the upper limit k of the search range.

In addition, the deterioration determining unit 120 performs least square regression on the set of Xc and Tc, and calculates the estimated slope a of the regression line. Specifically, the deterioration determination unit 120
a, b = lsq (Xc, Tc)
And slope a is calculated. Here, lsq () is a function that executes least square regression, and performs least square regression on a set of Xc and Tc, and determines the slope a and intercept b of the estimated regression line. Since the target slope is a negative value, an appropriate positive value may be set as the initial value (prev_a = 99).

  By substituting the slope a with R_p described above, the deterioration determination unit 120 allows the power of the straight line defined by the slope R_p and the fitting allowable width γ from the rear of the row of Xc and Tc, that is, from the long waiting time. Search for the measured value portion of the display latency that falls within the range. Specifically, the deterioration determination unit 120 executes the following procedure.

ここで、Xmin, Xmaxはそれぞれ、X2'の分布の裾をβ/2ずつ切り捨てた残りの最小値及び最大値であり、例えば、β=0.05 (=5%)である場合、Xmin, Xmaxはそれぞれ２．５パーセンタイル及び９７．５パーセンタイルとなる。また、if条件が成立した場合、(i-1)番目までは許容範囲内にあり、劣化判定部１２０は、劣化している範囲を示す劣化閾値として(i-1)番目の値を出力する。他方、if条件が成立しない場合、劣化判定部１２０は、カウンタiをインクリメントし、forループを繰り返す。

Here, Xmin and Xmax are the remaining minimum and maximum values obtained by rounding down the tail of the distribution of X2 ′ by β / 2, for example, when β = 0.05 (= 5%), Xmin and Xmax are The 2.5 percentile and 97.5 percentile respectively. Also, if the if condition is satisfied, the (i-1) th is within the allowable range, and the deterioration determining unit 120 outputs the (i-1) th value as the deterioration threshold indicating the deteriorated range. . On the other hand, when the if condition is not satisfied, the deterioration determination unit 120 increments the counter i and repeats the for loop.

  The deterioration determination unit 120 determines that a measurement value portion that matches a power line is detected when D is a value obtained by the above-described procedure and K> D and D> s, and the power value distribution is determined as a power distribution. It is determined that the waiting time of the target web page has deteriorated. In addition, the deterioration determination unit 120 outputs a deterioration threshold value T [Nr-D], and the measurement value portion is the (Nr-D) th and subsequent measurement values in the list T, and the deterioration threshold value T [Nr-D]. It can be determined that the above measured values are deteriorated.

  On the other hand, when D ≦ s, the deterioration determination unit 120 determines that there is no measurement value portion that applies to the power line in the measurement value group, and that the waiting time has not deteriorated.

  When k ≦ D, the deterioration determination unit 120 determines that the waiting time deterioration has occurred for a large number of measurement values. In this case, the deterioration determination unit 120 does not output the deterioration threshold because it cannot specify the deterioration threshold.

  The above-described power line fitting process is merely an example, and the power line fitting process according to the present invention is not limited to this, and any other appropriate value for detecting a measurement value portion that can be applied to the power line. Processing may be applied.

  Next, a waiting time evaluation process according to an embodiment of the present invention will be described with reference to FIG. The waiting time evaluation process is executed by the waiting time evaluation apparatus 100 described above. FIG. 3 is a flowchart showing the waiting time evaluation process according to one embodiment of the present invention.

  As illustrated in FIG. 3, in step S <b> 101, the waiting time evaluation apparatus 100 acquires a measured value group of display waiting time of the target web page. Specifically, the waiting time evaluation apparatus 100 receives, as input data, a sample group of measurement results of display waiting time required to display a web page to be evaluated on a terminal or the like, or displays a display waiting time. You may acquire the sample group of a measurement result by measuring by itself. For example, when N measurement values are acquired, the waiting time evaluation apparatus 100 forms a list of waiting time measurement values T = [t_1, t_2,..., T_N] by sorting the acquired measurement value groups in ascending order. .

  In step S <b> 102, the waiting time evaluation apparatus 100 determines whether there is a measured value portion that follows the power distribution in the acquired measured value group of display waiting time. Specifically, as described above with respect to the deterioration determination unit 120, the waiting time evaluation apparatus 100 performs a logarithmic transformation on the measurement value portion extracted from the measurement value group of the acquired display waiting time, thereby cumulative accumulation distribution. Data is generated, and it is determined whether or not the generated cumulative complementary distribution data can be applied to the power line. If it can be applied to the power line, it is determined that the extracted measurement value portion follows the power distribution.

  When the measured value portion according to the power distribution is detected (S102: Yes), in step S103, the waiting time evaluation apparatus 100 determines that the waiting time of the target web page has deteriorated. On the other hand, when the measured value portion according to the power distribution is not detected (S102: No), in step S104, the waiting time evaluation apparatus 100 determines that the waiting time of the target web page has not deteriorated.

  Next, with reference to FIG. 4, a hardware configuration of a waiting time evaluation apparatus according to an embodiment of the present invention will be described. FIG. 4 is a block diagram showing a hardware configuration of a waiting time evaluation apparatus according to an embodiment of the present invention.

  As shown in FIG. 4, the latency evaluation device 100 may typically be realized by a server, for example, a drive device 101, an auxiliary storage device 102, a memory device 103, which are interconnected via a bus, A processor 104, an interface device 105, and a communication device 106 are included. Various computer programs including the programs for realizing the various functions and processes described above in the prediction apparatus 100 are provided by a recording medium 107 such as a CD-ROM (Compact Disk-Read Only Memory), a DVD (Digital Versatile Disk), or a flash memory. May be. When the recording medium 107 storing the program is set in the drive device 101, the program is installed from the recording medium 107 to the auxiliary storage device 102 via the drive device 101. However, it is not always necessary to install the program using the recording medium 107, and the program may be downloaded from any external device via a network or the like. The auxiliary storage device 102 stores the installed program and also stores necessary files and data. The memory device 103 reads the program and data from the auxiliary storage device 102 and stores them when there is an instruction to start the program. The processor 104 executes various functions and processes of the waiting time evaluation apparatus 100 according to various data such as a program stored in the memory device 103 and parameters necessary for executing the program. The interface device 105 is used as a communication interface for connecting to a network or an external device. The communication device 106 executes various communication processes for communicating with a network such as the Internet. However, the hardware configuration described above is merely an example, and the waiting time evaluation apparatus 100 is not limited to the hardware configuration described above, and may be realized by any other appropriate hardware configuration.

  Note that each functional unit and each step of the waiting time evaluation apparatus 100 described above may be realized by the processor 104 executing a program stored in the memory device 103 of the computer.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to the specific embodiment mentioned above, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

100 wait time evaluation device 110 data acquisition unit 120 deterioration determination unit

Claims (6)

A waiting time evaluation device for evaluating the waiting time of a web page,
A data acquisition unit that acquires a measurement value group of display latency of the target web page;
A degradation determination unit that determines that the waiting time of the target web page has deteriorated when detecting a measured value portion that follows a power distribution in the acquired measurement value group of display waiting time;
I have a,
The deterioration determining unit generates cumulative complementary distribution data by performing logarithmic conversion on a measurement value portion extracted from the acquired measurement value group of display waiting time, and the generated cumulative complementary distribution data is a power line And if it can be applied to the power line, it is determined that the extracted measurement value portion follows a power distribution,
The data acquisition unit acquires the slope of the power line, the fitting allowable width, the fitting truncation width, the end point truncation threshold, the determination lower limit threshold, and the determination upper limit threshold,
The deterioration determination unit detects a measured value portion according to the power distribution based on the obtained slope of the straight line, allowable fitting width, fitting cut-off width, end-point cut-off threshold, determination lower-limit threshold, and determination upper-limit threshold. . A waiting time evaluation device for evaluating the waiting time of a web page,
A data acquisition unit that acquires a measurement value group of display latency of the target web page;
A degradation determination unit that determines that the waiting time of the target web page has deteriorated when detecting a measured value portion that follows a power distribution in the acquired measurement value group of display waiting time;
Have
The deterioration determining unit generates cumulative complementary distribution data by performing logarithmic conversion on a measurement value portion extracted from the acquired measurement value group of display waiting time, and the generated cumulative complementary distribution data is a power line And if it can be applied to the power line, it is determined that the extracted measurement value portion follows a power distribution,
The data acquisition unit acquires a fitting allowable width of the power line, a fitting cut-off width, an endpoint cut-off threshold, a determination lower limit threshold, and a determination upper limit threshold,
The deterioration determination unit calculates a slope of the power line by performing a least square regression on the cumulative complementary distribution data, and calculates the slope of the calculated power line, the allowable width of fitting of the acquired power line, and a fit. truncation width, endpoint truncation threshold, based on the determination lower limit threshold and the determination upper limit threshold value, waiting time evaluation device you detect measurement portion according the powers distribution.   The waiting time evaluation apparatus according to claim 1, wherein when the deterioration determining unit determines that the waiting time of the target web page is deteriorated, the deterioration determining unit outputs a deterioration threshold indicating a deteriorated range. A latency evaluation method executed by a device for evaluating a latency of a web page,
Obtaining a measurement value group of display latency of the target web page;
Determining whether or not there is a measurement value portion that follows a power distribution in the measurement value group of the acquired display waiting time; and
When detecting a measured value portion according to the power distribution, determining that the waiting time of the target web page has deteriorated;
I have a,
The determining step generates cumulative complementary distribution data by performing logarithmic conversion on the measurement value portion extracted from the acquired display latency measurement value group, and the generated cumulative complementary distribution data is a straight line And if it can be applied to the power line, it is determined that the extracted measurement value portion follows a power distribution,
The obtaining step obtains the slope of the power line, the fitting allowable width, the fitting truncation width, the end point truncation threshold value, the determination lower limit threshold value, and the determination upper limit threshold value,
The determining step includes a waiting time evaluation method for detecting a measured value portion according to the power distribution based on the obtained slope of the straight line, allowable fitting width, fitting cut-off width, end-point cut-off threshold, determination lower-limit threshold, and determination upper-limit threshold. . A latency evaluation method executed by a device for evaluating a latency of a web page,
  Obtaining a measurement value group of display latency of the target web page;
  Determining whether or not there is a measurement value portion that follows a power distribution in the measurement value group of the acquired display waiting time; and
  When detecting a measured value portion according to the power distribution, determining that the waiting time of the target web page has deteriorated;
Have
  The determining step generates cumulative complementary distribution data by performing logarithmic conversion on the measurement value portion extracted from the acquired display latency measurement value group, and the generated cumulative complementary distribution data is a straight line And if it can be applied to the power line, it is determined that the extracted measurement value portion follows a power distribution,
  The obtaining step obtains an allowable width of fitting of the power line, a fitting cut-off width, an end-point cut-off threshold, a determination lower limit threshold, and a determination upper limit threshold;
  The determining step calculates a slope of the power line by performing a least square regression on the cumulative complementary distribution data, and calculates the slope of the power line, the allowable fit width of the acquired power line, and a fit. A waiting time evaluation method for detecting a measured value portion according to the power distribution based on a cut-off width, an end-point cut-off threshold, a determination lower limit threshold, and a determination upper limit threshold. 6. The waiting time evaluation method according to claim 4 or 5, wherein when the determining step determines that the waiting time of the target web page is deteriorated, a deterioration threshold value indicating a deteriorated range is output.

Images