WO2019033677A1 - 确定用户行为衰退倾向的方法、装置及电子设备 - Google Patents

确定用户行为衰退倾向的方法、装置及电子设备 Download PDF

Info

Publication number
WO2019033677A1
WO2019033677A1 PCT/CN2017/118773 CN2017118773W WO2019033677A1 WO 2019033677 A1 WO2019033677 A1 WO 2019033677A1 CN 2017118773 W CN2017118773 W CN 2017118773W WO 2019033677 A1 WO2019033677 A1 WO 2019033677A1
Authority
WO
WIPO (PCT)
Prior art keywords
tendency
predetermined
user behavior
recession
user
Prior art date
Application number
PCT/CN2017/118773
Other languages
English (en)
French (fr)
Inventor
谢本银
殷良鹰
Original Assignee
北京小度信息科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 北京小度信息科技有限公司 filed Critical 北京小度信息科技有限公司
Publication of WO2019033677A1 publication Critical patent/WO2019033677A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data

Definitions

  • the present disclosure relates to the field of computer technologies, and in particular, to a method, an apparatus, an electronic device, and a computer storage medium for determining a tendency of a user to decline in behavior.
  • APP Application, referred to as APP.
  • Each APP has a certain number of users, and these users may experience one or more stages of new user period, rising period, stable period, recession period, and drain period during the use of the APP.
  • the period in which the amount of user orders is gradually reduced is called the recession period of user behavior, and the degree of decline in user behavior is called the tendency of decline.
  • Embodiments of the present disclosure provide a method, apparatus, electronic device, and computer storage medium for determining a tendency of user behavior to decline.
  • a method of determining a tendency of a user to decline in behavior is provided in an embodiment of the present disclosure.
  • methods for determining the tendency of user behavior to decline include:
  • a total decline tendency value of the predetermined user behavior is determined based on the first recession tendency factor and the second recession tendency factor.
  • the disclosure determines the first recession tendency factor according to the number of times the predetermined user behavior is generated, including:
  • the first decay tendency factor is determined based on a change in the average number of times the predetermined user behavior is generated within a predetermined time period.
  • the first decay tendency factor is calculated as follows:
  • GF is a first recession tendency factor
  • i is a first predetermined time period
  • j is a second predetermined time period
  • R i is an average number of times of generating a predetermined user behavior in a first predetermined time period i
  • R j is a second predetermined The average number of times of the predetermined user behavior is generated in the time period j, wherein the time starting point of the first predetermined time period i is earlier than the second predetermined time period j, a is the base number, n is the number threshold, and the value is greater than or equal to 0. Integer.
  • the disclosure determines the second recession tendency factor according to the time interval at which the predetermined user behavior is generated, including:
  • the first time interval is a time interval at which a predetermined user behavior is generated from the current time
  • the second time interval is a predetermined time period in the third predetermined time period.
  • the average time interval of user behavior, the end of the third predetermined time period is before the time when the predetermined user behavior was last generated.
  • the second recession tendency factor is calculated as follows:
  • R represents the first time interval
  • Mi represents the second time interval
  • the end of the third predetermined time period precedes the time when the predetermined user behavior was last generated.
  • the total decline propensity value of the predetermined user behavior is calculated as follows:
  • DI is the total decline propensity value of the predetermined user behavior
  • GF is the first recession tendency factor
  • GR is the second recession tendency factor
  • b is the weight value
  • the method further includes:
  • the user whose total decline tendency value of the predetermined user behavior is less than the first predetermined value is output according to the level of the total recession tendency value of the predetermined user behavior.
  • an embodiment of the present disclosure provides an apparatus for determining a tendency of a user to decline in behavior, including:
  • a first determining module configured to determine a first recession tendency factor according to a number of times the predetermined user behavior is generated
  • a second determining module configured to determine a second recession tendency factor according to a time interval at which a predetermined user behavior is generated
  • the third determining module is configured to determine a total recession tendency value of the predetermined user behavior based on the first recession tendency factor and the second recession tendency factor.
  • the first determining module includes:
  • the first determining sub-module is configured to determine a first decay tendency factor based on a change in the average number of times the predetermined user behavior is generated within a predetermined time period.
  • the first recession tendency factor is calculated as follows:
  • GF is a first recession tendency factor
  • i is a first predetermined time period
  • j is a second predetermined time period
  • R i is an average number of times of generating a predetermined user behavior in a first predetermined time period i
  • R j is a second predetermined The average number of times of the predetermined user behavior is generated in the time period j, wherein the time starting point of the first predetermined time period i is earlier than the second predetermined time period j, a is the base number, n is the number threshold, and the value is greater than or equal to 0. Integer.
  • the second determining module includes:
  • a second determining submodule configured to determine a second recession tendency factor according to a ratio of the first time interval and the second time interval, where the first time interval is a time interval during which the last user behavior is generated from the current time, and the second time interval An average time interval for generating a predetermined user behavior for a third predetermined time period, the end of the third predetermined time period being before the time when the predetermined user behavior was last generated.
  • the second recession tendency factor is calculated as follows:
  • R represents the first time interval
  • Mi represents the second time interval
  • the end of the third predetermined time period precedes the time when the predetermined user behavior was last generated.
  • the total decline propensity value of the predetermined user behavior is calculated as follows:
  • DI is the total decline propensity value of the predetermined user behavior
  • GF is the first recession tendency factor
  • GR is the second recession tendency factor
  • b is the weight value
  • the device further includes:
  • the output module is configured to output a user whose total decline tendency value of the predetermined user behavior is less than the first predetermined value according to the level of the total decline tendency value of the predetermined user behavior.
  • the functions can be implemented in hardware or in hardware by executing the corresponding software.
  • the hardware or software includes one or more modules corresponding to the functions described above.
  • the structure of the device for determining the tendency of user behavior to decline includes a memory and a processor for storing one or more devices supporting the determination of a predetermined user behavior degradation tendency to perform the determining user behavior in the first aspect above.
  • a computer instruction of a method of recession propensity the processor being configured to execute computer instructions stored in the memory.
  • the means for determining the tendency of the user to decline in behavior may also include a communication interface, the means for determining the tendency of the user to decline in behavior to communicate with other devices or communication networks.
  • an embodiment of the present disclosure further provides an electronic device including a memory and a processor; wherein the memory is configured to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor A method of determining the tendency of user behavior to decline in the first aspect.
  • an embodiment of the present disclosure provides a computer readable storage medium for storing computer instructions for determining a user behavior degradation tendency, the method comprising the method for determining a user behavior degradation tendency in the first aspect described above Computer instructions involved.
  • the embodiment of the present disclosure determines a first recession tendency factor according to the number of times the predetermined user behavior is generated, and determines a second recession tendency factor according to a time interval at which the predetermined user behavior is generated, and comprehensively considers the first recession tendency factor and the second recession tendency factor to determine the predetermined schedule.
  • the total decline propensity value of user behavior uses the first degradation factor to reduce the frequency of the user's predetermined user behavior as a reference indicator for the user's degree of decline, and can quickly acquire the user in the system platform in a recession period, and further adopt the second regression factor to further the user.
  • the ratio of the time interval in which the predetermined user behavior is not generated relative to the time interval in which the predetermined user behavior was previously generated is another reference indicator for characterizing the degree of user decline, compared to the single time interval in the prior art based on the user's recent failure to generate a predetermined behavior. It is more comprehensive, and by combining the above two factors, it can more fully and accurately determine the decline tendency of user behavior, so that timely intervention measures can greatly reduce the loss of the application platform.
  • FIG. 1 shows a flow chart of a method of determining a tendency of a user to decline in behavior according to an embodiment of the present disclosure
  • FIG. 2 is a block diagram showing the structure of a device for determining user behavior degradation tendency according to an embodiment of the present disclosure
  • FIG. 3 is a block diagram of an electronic device suitable for implementing a method of determining a user behavior degradation tendency in accordance with an embodiment of the present disclosure.
  • the prior art characterization of the platform user's tendency to decline often only considers the time when the predetermined user behavior is generated. The user does not make targeted discrimination.
  • the characteristics of the recession tendency are very simple, and the depiction of the recession tendency is not accurate enough. For example, for users who have different time intervals for generating predetermined user behaviors, when their time intervals for not generating predetermined user behaviors are the same, the degraded tendency values of the two obtained by the prior art are the same, which is obviously not objective.
  • the first recession tendency factor is determined according to the number of times the predetermined user behavior is generated, and the second recession tendency factor is determined according to the time interval at which the predetermined user behavior is generated, and the first recessive tendency factor and the second recessive tendency factor are comprehensively considered.
  • the first degradation factor is used to reduce the frequency of the user's predetermined user behavior as a reference indicator for the user's degree of decline. The user in the system platform can be obtained as soon as possible, and the user is not yet generated by the second recession factor.
  • the ratio of the time interval of the behavior to the time interval in which the predetermined user behavior was previously generated is another reference indicator for the degree of user decay, which is more comprehensive than the single time interval in the prior art based on the user's recent failure to generate a predetermined behavior.
  • FIG. 1 illustrates a flow chart of a method of determining a tendency of a user to decline in behavior in accordance with an embodiment of the present disclosure.
  • the method for determining the tendency of user behavior to decline includes the following steps S101-S103:
  • step S101 determining a first recession tendency factor according to the number of times the predetermined user behavior is generated
  • step S102 determining a second recession tendency factor according to a time interval at which a predetermined user behavior is generated
  • step S103 a total recession tendency value of the predetermined user behavior is determined based on the first recession tendency factor and the second recession tendency factor.
  • the user behavior may be an effective behavior of the user in the system, including various operations of the user on the system object, for example, for the commercial operation application platform, the user uses the application platform to perform the user behavior of the order.
  • the predetermined user behavior is a predefined user behavior by which the application platform wants to measure the user's tendency to decline. For example, for a take-out application platform, the user behavior that measures the user's tendency to decline may be the user's order behavior, and for a product marketing website. Application platform, the user behavior of horizontal user decline tends to be the best user behavior may be the user's browsing behavior.
  • the predetermined user behavior may be different according to the content involved in the application platform, and is specifically set according to actual conditions.
  • the period in which the number of times a user generates a predetermined user behavior during a certain period of time may be referred to as a recession period, and the degree to which the user's predetermined user behavior is currently degraded is referred to as a recession tendency.
  • a recession period the number of times a user places an order during a certain period of time is reduced, and this period is a recession period in which the user places an order, and the degree of the number of times the user places an order is a tendency for the user to place a decline.
  • the total decline tendency value of a certain predetermined user behavior of the user is determined by comprehensively considering the number of times the user generates the predetermined user behavior and the time interval. For example, for a commercial operation application platform, when the number of times an order is placed by a user in a certain period of time is reduced, and the time interval of placing an order is also long, the user may determine the user during the period by the total decline tendency value of the order placed by the user. There is a tendency to decline, and the commercial operation platform intervenes through various measures to prevent user loss in the case of timely capturing the tendency of users to decline.
  • the above method for determining the user's tendency to decline is not only applicable to the commercial operation application platform, but also applicable to various application platforms for users.
  • the application platform can feedback the user to the product involved in the application platform by determining the user's decline tendency. Or the attention and use of the service, and further improve the quality of the product or service provided by the application platform according to the feedback information, etc., can promote the further development of the application platform and improve the user experience.
  • step S101 that is, determining the first recession tendency factor according to the number of times the predetermined user behavior is generated, further comprising the step of: generating an average number of predetermined user behaviors according to the predetermined time period.
  • the change determines the first recession propensity factor.
  • the first fading factor may be determined by the number of times the user generates the predetermined user behavior in a period of time. According to the chronological order, if the number of times is gradually decreased, it may be inferred that the predetermined user behavior generated by the user is decreasing. This can reflect the user's tendency to decline to a certain extent.
  • the total number of orders placed 7 days in the previous week is Ri
  • the total number of orders placed in the last 7 days is Rj
  • Ri>Rj at which time the user can initially determine the order placed.
  • Behavior is in a recession, users have a tendency to decline, and the magnitude of the recession tends to be related to the absolute value of the difference between Ri and Rj. If Ri>Rj, and the absolute value of the difference between Ri and Rj is large, the user's tendency to decline is large, and the absolute value of the difference is small, indicating that the user's tendency to decline is small.
  • the high-value users of the platform can be obtained as soon as possible, so that timely intervention measures can be taken according to the tendency of the user behavior to decline, which can greatly Reduce the loss of the application platform.
  • the first decay tendency factor can be calculated as follows:
  • GF is a first recession tendency factor
  • i is a first predetermined time period
  • j is a second predetermined time period
  • R i is an average number of times of generating a predetermined user behavior in a first predetermined time period i
  • R j is a second predetermined The average number of times of the predetermined user behavior is generated in the time period j, wherein the time starting point of the first predetermined time period i is earlier than the second predetermined time period j, a is the base number, n is the number threshold, and the value is greater than or equal to 0. Integer.
  • the time start of the first predetermined time period i is earlier than the second predetermined time period j
  • the time end point of the first predetermined time period i may be earlier than the second predetermined time period j, or may be the same.
  • the first predetermined time period i is In the last week
  • the second predetermined time period j is the previous week, etc.
  • the first predetermined time period i and the first predetermined time period j may be selected according to actual conditions, as long as R i and R j can be made to show the user from the last time.
  • the second predetermined time period j is the latest week
  • R j is the average order number within the last 7 days of the last week
  • the first predetermined time period i is the latest month
  • R i is the most recent month.
  • an exponential function is used to quantify a downward trend of the number of times the user generates a predetermined user behavior, and an average number of times the predetermined user behavior is generated in the first predetermined time period is greater than an average number of times the predetermined user behavior is generated in the second predetermined time period.
  • the exponential function with the base a is used to quantify the downward trend of the predetermined user behavior, and The average number of times the predetermined user behavior is generated within a predetermined time period is less than or equal to the sum of the average number of times the predetermined user behavior is generated and the number of times threshold n in the second predetermined time period, that is, the average number of times the predetermined user behavior is generated in the near future is compared to the previous period. If the statement is not reduced, or if the reduction is still within the range of the threshold n, then the first decay tendency factor is directly set to zero.
  • the values of parameters such as i, j, a, and n can be set according to actual conditions.
  • the users who have not placed orders for 40 consecutive days are basically regarded as the loss, so the tendency of the decline is only to describe the users who have not placed the order for more than 40 days, so the above parameters can be determined according to the following principles. select:
  • the first recession tendency factor focuses on the decrease of the single quantity, and if the average order quantity of the user on the 7th is considered to be greater than 2, the value of the user who has a large amount of decline is high, so the total The decline propensity value can be dominated by the value of the first decay tendency factor.
  • the minimum value of the first recession tendency factor is greater than or equal to the maximum value of the second recessive tendency factor.
  • the value of the base a is determined.
  • the value of the base a can also be adjusted according to other cases, and the embodiment does not impose any restrictions. .
  • the number threshold n can be an integer greater than or equal to 0, which may be determined according to actual conditions, and may also be adjusted to different values in different periods, for example, when determining the tendency of the user to decline in order behavior, the value of n may be adjusted during the off-season. Set to a larger value, and set the value of n to a smaller value during the promotion season, because it is normal for the low-season users to place orders less than the number of orders in the promotion season, so it can be adjusted by adjusting the threshold n. Objectively portray the tendency of users to decline.
  • the calculation formula of the first recession tendency factor in the alternative implementation adopts an exponential function, so that as the number of times the user generates a predetermined user behavior decreases, the value of the first recession tendency factor increases, so that the attenuation tends to grow faster. Because the user who generates a high number of predetermined user behaviors is usually a high-value user of the application platform, the tendency of such users to decline requires more timely intervention, and the first decline factor in the optional implementation can timely monitor the user to generate a predetermined user. The area of decline in behavior and timely intervention.
  • the step of determining the second recession tendency factor according to the time interval for generating the predetermined user behavior in step 102 further includes the following steps: according to the first time interval and the second time The ratio of the intervals determines a second recession tendency factor, the first time interval is a time interval during which the predetermined user behavior is generated from the current time, and the second time interval is an average time interval for generating a predetermined user behavior in the third predetermined time period, and the third time interval The end of the predetermined time period is before the time when the predetermined user behavior was last generated.
  • the second recession tendency factor is determined by the ratio of the time interval between the recent generation of the predetermined user behavior from the current time and the average time interval in which the predetermined user behavior was previously generated, and if the ratio is larger, the second recession tendency factor The larger, the greater the user's tendency to decline.
  • the first time interval may be determined by calculating the time difference between the current time and the last time the user generated the predetermined behavior, and the unit of measurement may be days, hours, and the like.
  • the system platform records that the time when the user finally generated the predetermined behavior is 4 days ago, and the time interval from the current time is 4 days, and the predetermined user behavior is generated within a period of 4 days ago (such as 7 days or half a month, etc.).
  • the average time interval is 2 days, and the ratio of the two is 2, and the magnitude of the second recession tendency factor is determined by the ratio 2.
  • the second decay tendency factor is calculated as follows:
  • R represents a first time interval, that is, a time interval at which a predetermined user behavior is generated from the current time
  • Mi represents a second time interval, that is, an average time interval at which a predetermined user behavior is generated in a third predetermined time period, where m is time Interval threshold.
  • the last time interval between the second time interval and the time interval threshold is generated when the first time interval of the predetermined user behavior from the current time is less than or equal to the average time interval for generating the predetermined user behavior in the third predetermined time period.
  • the second recession tendency factor is determined by the logarithmic function described above.
  • the values of the third predetermined time period, m and other parameters can be set according to actual conditions. If, for the commercial operation application platform of the take-out type, the user who has not placed the order for 40 consecutive days is basically regarded as the loss, the third predetermined time period can be set to 40, and the Mi is the average of 40 days before the last time the user places the order. Order time interval.
  • the time interval threshold m may take an integer greater than or equal to 0, which may be determined according to actual conditions, and may be adjusted to different values at different times. For example, when determining the tendency of the user to decline in order behavior, the promotion season may be m.
  • the value of the value is set to a larger value, and the value of m is set to a smaller value during the promotion season, because the time interval for the promotion of the off-season user is longer than the time interval for placing the order during the promotion season, so the time interval can be passed.
  • the adjustment of the threshold m can more objectively portray the user's tendency to decline.
  • the total decay propensity value for the predetermined user behavior is calculated as follows:
  • DI is the total decline tendency value of the predetermined user behavior
  • GF is the first recession tendency factor
  • GR is the second recession tendency factor
  • b and c are the weight values.
  • the total decay tendency value is determined by weighting the first decay tendency factor and the second decay tendency factor.
  • the value of c can be set according to the actual situation, the influence of the first recession tendency factor and the second recession tendency factor in different industries and different situations can be different, and some industry first recession tendency factors can better reflect the user's recession. There is a tendency that some industries' second recession tendency factors may better reflect the user's tendency to decline. The proportion of the first recession tendency factor and the second recession tendency factor in the same industry may be different in the user's recession tendency. Determine the actual situation and set the values of parameters b and c.
  • the method further includes the step of outputting a user whose total decline tendency value of the predetermined user behavior is less than the first predetermined value according to the level of the total decline tendency value of the predetermined user behavior.
  • the user after determining the total decline tendency value of the user, if the total decline tendency value is less than the first predetermined value (which may be set according to actual conditions), the user is in a recession period, and corresponding intervention measures may be taken. To prevent the loss of users. By outputting all users whose total recession tendency value is less than the first predetermined value in order, and adopting different intervention strategies for different users or user groups, the loss of the application platform users can be prevented.
  • the platform should select the first decay factor, the second decay factor, and the total decline tendency value to calculate the total decline tendency value of the user, and the total decline tendency value is higher than the first value.
  • the user of the predetermined value conducts marketing intervention.
  • the six-week unrecovered user rate k can be used as the evaluation index, and the specific calculation formula is as follows:
  • the following is the data obtained from the test of a commercial operation application platform.
  • the following three tables are the percentage of the total decline tendency value obtained by xx, m1, m+1, and m+2, respectively.
  • the data and the corresponding k-value data are as follows:
  • the first column in Table 1 to Table 3 is the percentage data of the total fading tendency value of the user's ordering behavior by size
  • the second column is the number of users under the corresponding percentage
  • the third column is the k value of these users.
  • the apparatus for determining a user behavior degradation tendency includes a first determining module 201, a second determining module 202, and a third determining module 203:
  • the first determining module 201 is configured to determine a first recession tendency factor according to the number of times the predetermined user behavior is generated;
  • the second determining module 202 is configured to determine a second recession tendency factor according to a time interval at which a predetermined user behavior is generated;
  • the third determining module 203 is configured to determine a total recession tendency value of the predetermined user behavior based on the first recession tendency factor and the second recession tendency factor.
  • the user behavior may be an effective behavior of the user in the system, including various operations of the user on the system object, for example, for the commercial operation application platform, the user uses the application platform to perform the user behavior of the order.
  • the predetermined user behavior is a predefined user behavior by which the application platform wants to measure the user's tendency to decline. For example, for a take-out application platform, the user behavior that measures the user's tendency to decline may be the user's order behavior, and for a product marketing website. Application platform, the user behavior of horizontal user decline tends to be the best user behavior may be the user's browsing behavior.
  • the predetermined user behavior may be different according to the content involved in the application platform, and is specifically set according to actual conditions.
  • the period in which the number of times a user generates a predetermined user behavior during a certain period of time may be referred to as a recession period, and the degree to which the user's predetermined user behavior is currently degraded is referred to as a recession tendency.
  • a recession period the number of times a user places an order during a certain period of time is reduced, and this period is a recession period in which the user places an order, and the degree of the number of times the user places an order is a tendency for the user to place a decline.
  • the total decline tendency value of a certain predetermined user behavior of the user is determined by comprehensively considering the number of times the user generates the predetermined user behavior and the time interval. For example, for a commercial operation application platform, when the number of times an order is placed by a user in a certain period of time is reduced, and the time interval of placing an order is also long, the user may determine the user during the period by the total decline tendency value of the order placed by the user. There is a tendency to decline, and the commercial operation platform intervenes through various measures to prevent user loss in the case of timely capturing the tendency of users to decline.
  • the above method for determining the user's tendency to decline is not only applicable to the commercial operation application platform, but also applicable to various application platforms for users.
  • the application platform can feedback the user to the product involved in the application platform by determining the user's decline tendency. Or the attention and use of the service, and further improve the quality of the product or service provided by the application platform according to the feedback information, etc., can promote the further development of the application platform and improve the user experience.
  • the first determining module 201 includes: a first determining submodule configured to determine a first recession tendency factor according to a change in an average number of times the predetermined user behavior is generated within a predetermined time period.
  • the first fading factor may be determined by the number of times the user generates the predetermined user behavior in a period of time. According to the chronological order, if the number of times is gradually decreased, it may be inferred that the predetermined user behavior generated by the user is decreasing. This can reflect the user's tendency to decline to a certain extent.
  • the total number of orders placed 7 days in the previous week is Ri
  • the total number of orders placed in the last 7 days is Rj
  • Ri>Rj at which time the user can initially determine the order placed.
  • Behavior is in a recession, users have a tendency to decline, and the magnitude of the recession tends to be related to the absolute value of the difference between Ri and Rj. If Ri>Rj, and the absolute value of the difference between Ri and Rj is large, the user's tendency to decline is large, and the absolute value of the difference is small, indicating that the user's tendency to decline is small.
  • the high-value users of the platform can be obtained as soon as possible, so that timely intervention measures can be taken according to the tendency of the user behavior to decline, which can greatly Reduce the loss of the application platform.
  • the first decay tendency factor can be calculated as follows:
  • GF is a first recession tendency factor
  • i is a first predetermined time period
  • j is a second predetermined time period
  • R i is an average number of times of generating a predetermined user behavior in a first predetermined time period i
  • R j is a second predetermined The average number of times of the predetermined user behavior is generated in the time period j, wherein the time starting point of the first predetermined time period i is earlier than the second predetermined time period j, a is the base number, n is the number threshold, and the value is greater than or equal to 0. Integer.
  • the time start of the first predetermined time period i is earlier than the second predetermined time period j
  • the time end point of the first predetermined time period i may be earlier than the second predetermined time period j, or may be the same.
  • the first predetermined time period i is In the last week
  • the second predetermined time period j is the previous week, etc.
  • the first predetermined time period i and the first predetermined time period j may be selected according to actual conditions, as long as R i and R j can be made to show the user from the last time.
  • the second predetermined time period j is the latest week
  • R j is the average order number within the last 7 days of the last week
  • the first predetermined time period i is the latest month
  • R i is the most recent month.
  • an exponential function is used to quantify a downward trend of the number of times the user generates a predetermined user behavior, and an average number of times the predetermined user behavior is generated in the first predetermined time period is greater than an average number of times the predetermined user behavior is generated in the second predetermined time period.
  • the exponential function with the base a is used to quantify the downward trend of the predetermined user behavior, and The average number of times the predetermined user behavior is generated within a predetermined time period is less than or equal to the sum of the average number of times the predetermined user behavior is generated and the number of times threshold n in the second predetermined time period, that is, the average number of times the predetermined user behavior is generated in the near future is compared to the previous period. If the statement is not reduced, or if the reduction is still within the range of the threshold n, then the first decay tendency factor is directly set to zero.
  • the values of parameters such as i, j, a, and n can be set according to actual conditions.
  • the users who have not placed orders for 40 consecutive days are basically regarded as the loss, so the tendency of the decline is only to describe the users who have not placed the order for more than 40 days, so the above parameters can be determined according to the following principles. select:
  • the first recession tendency factor focuses on the decrease of the single quantity, and if the average order quantity of the user on the 7th is considered to be greater than 2, the value of the user who has a large amount of decline is high, so the total The decline propensity value can be dominated by the value of the first decay tendency factor.
  • the minimum value of the first recession tendency factor is greater than or equal to the maximum value of the second recessive tendency factor.
  • the value of the base a is determined.
  • the value of the base a can also be adjusted according to other cases, and the embodiment does not impose any restrictions. .
  • the number threshold n can be an integer greater than or equal to 0, which may be determined according to actual conditions, and may also be adjusted to different values in different periods, for example, when determining the tendency of the user to decline in order behavior, the value of n may be adjusted during the off-season. Set to a larger value, and set the value of n to a smaller value during the promotion season, because it is normal for the low-season users to place orders less than the number of orders in the promotion season, so it can be adjusted by adjusting the threshold n. Objectively portray the tendency of users to decline.
  • the calculation formula of the first recession tendency factor in the alternative implementation adopts an exponential function, so that as the number of times the user generates a predetermined user behavior decreases, the value of the first recession tendency factor increases, so that the attenuation tends to grow faster. Because the user who generates a high number of predetermined user behaviors is usually a high-value user of the application platform, the tendency of such users to decline requires more timely intervention, and the first decline factor in the optional implementation can timely monitor the user to generate a predetermined user. The area of decline in behavior and timely intervention.
  • the second determining module 202 includes: a second determining submodule, determining a second recession tendency factor according to a ratio of the first time interval and the second time interval, where the first time interval is The time interval between the predetermined user behavior and the current time is generated for the last time, and the second time interval is an average time interval for generating a predetermined user behavior in the third predetermined time period, and the end of the third predetermined time period is before the time when the predetermined user behavior is last generated.
  • the second recession tendency factor is determined by the ratio of the time interval between the recent generation of the predetermined user behavior from the current time and the average time interval in which the predetermined user behavior was previously generated, and if the ratio is larger, the second recession tendency factor The larger, the greater the user's tendency to decline.
  • the first time interval may be determined by calculating the time difference between the current time and the last time the user generated the predetermined behavior, and the unit of measurement may be days, hours, and the like.
  • the system platform records that the time when the user finally generated the predetermined behavior is 4 days ago, and the time interval from the current time is 4 days, and the predetermined user behavior is generated within a period of 4 days ago (such as 7 days or half a month, etc.).
  • the average time interval is 2 days, and the ratio of the two is 2, and the magnitude of the second recession tendency factor is determined by the ratio 2.
  • the second decay tendency factor is calculated as follows:
  • R represents a first time interval, that is, a time interval at which a predetermined user behavior is generated from the current time
  • Mi represents a second time interval, that is, an average time interval at which a predetermined user behavior is generated in a third predetermined time period, where m is time Interval threshold.
  • the last time interval between the second time interval and the time interval threshold is generated when the first time interval of the predetermined user behavior from the current time is less than or equal to the average time interval for generating the predetermined user behavior in the third predetermined time period.
  • the time interval during which the user generates the predetermined user behavior increases, and when the time interval outside the preset time interval threshold m is exceeded, the user is in a recession period, and thus the second decay tendency factor is determined by the logarithmic function described above.
  • the values of the third predetermined time period, m and other parameters can be set according to actual conditions. If, for the commercial operation application platform of the take-out type, the user who has not placed the order for 40 consecutive days is basically regarded as the loss, the third predetermined time period can be set to 40, and the Mi is the average of 40 days before the last time the user places the order. Order time interval.
  • the time interval threshold m may take an integer greater than or equal to 0, which may be determined according to actual conditions, and may be adjusted to different values at different times. For example, when determining the tendency of the user to decline in order behavior, the promotion season may be m.
  • the value of the value is set to a larger value, and the value of m is set to a smaller value during the promotion season, because the time interval for the promotion of the off-season user is longer than the time interval for placing the order during the promotion season, so the time interval can be passed.
  • the adjustment of the threshold m can more objectively portray the user's tendency to decline.
  • the total decay propensity value of the predetermined user behavior is calculated as follows:
  • DI is the total decline tendency value of the predetermined user behavior
  • GF is the first recession tendency factor
  • GR is the second recession tendency factor
  • b and c are the weight values.
  • the total decay tendency value is determined by weighting the first decay tendency factor and the second decay tendency factor.
  • the value of c can be set according to the actual situation, the influence of the first recession tendency factor and the second recession tendency factor in different industries and different situations can be different, and some industry first recession tendency factors can better reflect the user's recession. There is a tendency that some industries' second recession tendency factors may better reflect the user's tendency to decline. The proportion of the first recession tendency factor and the second recession tendency factor in the same industry may be different in the user's recession tendency. Determine the actual situation and set the values of parameters b and c.
  • the apparatus further includes: an output module configured to output a total decline tendency value of the predetermined user behavior that is less than the first predetermined value according to a level of a total decline tendency value of the predetermined user behavior user.
  • an output module configured to output a total decline tendency value of the predetermined user behavior that is less than the first predetermined value according to a level of a total decline tendency value of the predetermined user behavior user.
  • the above-described apparatus for determining the tendency of user behavior to decline also includes some other well-known structures, such as processors, memories, etc., in order to unnecessarily obscure the embodiments of the present disclosure, these well-known structures are in FIG. Not shown.
  • FIG. 3 is a block diagram of an electronic device suitable for implementing a method of determining a tendency to decline in user behavior in accordance with an embodiment of the present disclosure.
  • the electronic device 300 includes a central processing unit (CPU) 301 that can be loaded into a program in a random access memory (RAM) 303 according to a program stored in a read only memory (ROM) 302 or a program stored from the storage portion 308.
  • CPU central processing unit
  • RAM random access memory
  • ROM read only memory
  • the various processes in the embodiment shown in Fig. 1 described above are executed.
  • RAM 303 various programs and data required for the operation of the system 300 are also stored.
  • the CPU 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304.
  • An input/output (I/O) interface 305 is also coupled to bus 304.
  • the following components are connected to the I/O interface 305: an input portion 306 including a keyboard, a mouse, etc.; an output portion 307 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), and the like, and a storage portion 308 including a hard disk or the like. And a communication portion 309 including a network interface card such as a LAN card, a modem, or the like. The communication section 309 performs communication processing via a network such as the Internet.
  • Driver 310 is also connected to I/O interface 305 as needed.
  • a removable medium 311 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory or the like is mounted on the drive 310 as needed so that a computer program read therefrom is installed into the storage portion 308 as needed.
  • an embodiment of the present disclosure includes a computer program product comprising a computer program tangibly embodied on a readable medium therewith, the computer program comprising program code for performing the key value data processing method of FIG.
  • the computer program can be downloaded and installed from the network via the communication portion 309, and/or installed from the removable medium 311.
  • each block in the roadmap or block diagram can represent a module, a program segment, or a portion of code that contains one or more executables for implementing the specified logical functions. instruction.
  • the functions noted in the blocks may also be produced in a different order than that illustrated in the drawings. For example, two successively represented blocks may in fact be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts can be implemented in a dedicated hardware-based system that performs the specified function or operation. Or it can be implemented by a combination of dedicated hardware and computer instructions.
  • the units or modules described in the embodiments of the present disclosure may be implemented by software or by hardware.
  • the described units or modules may also be provided in a processor, the names of which do not in any way constitute a limitation of the unit or module itself.
  • the present disclosure further provides a computer readable storage medium, which may be a computer readable storage medium included in the apparatus in the above embodiment; or may exist separately, not assembled A computer readable storage medium in a device.
  • a computer readable storage medium stores one or more programs that are used by one or more processors to perform the methods described in the present disclosure.

Landscapes

  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Telephonic Communication Services (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种确定用户衰退倾向的方法、装置及电子设备。方法包括：根据产生预定用户行为的次数确定第一衰退倾向因子(S101)；根据产生预定用户行为的时间间隔确定第二衰退倾向因子(S102)；基于第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值(S103)。能够尽快获取到系统平台中处于衰退期的用户，相较于已有技术中单一的根据用户近期未产生预定行为的时间间隔来说更加全面，并且通过综合以上两种因子，能够更加全面准确的确定用户行为的衰退倾向，从而采取及时干预措施，能大大减少应用平台的损失。

Description

确定用户行为衰退倾向的方法、装置及电子设备 技术领域
本公开涉及计算机技术领域，具体涉及一种确定用户行为衰退倾向的方法、装置、电子设备及计算机存储介质。
背景技术
互联网技术的不断发展已经大大改变了人们的生活方式，比如人们的出行方式、购物方式、配送方式等都在发生巨大变化，为满足用户的各种需求，开发出了各种相应的应用程序(Application，简称APP)。每一个APP拥有一定的用户，这些用户在APP的使用过程可能会经历新用户期、上升期、稳定期、衰退期、流失期等一个或者多个阶段。用户下单量逐渐减少的时期称为用户行为的衰退期，用户行为衰退的程度称为衰退倾向。
发明内容
本公开实施例提供一种确定用户行为衰退倾向的方法、装置、电子设备及计算机存储介质。
第一方面，本公开实施例中提供了一种确定用户行为衰退倾向的方法。
具体的，确定用户行为衰退倾向的方法，包括：
根据产生预定用户行为的次数确定第一衰退倾向因子；
根据产生预定用户行为的时间间隔确定第二衰退倾向因子；
基于第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值。
结合第一方面，本公开在第一方面的第一种实现方式中，根据产生预定用户行为的次数确定第一衰退倾向因子，包括：
根据在预定时间周期内产生预定用户行为的平均次数变化确定第一衰退倾向因子。
结合第一方面、第一方面的第一种实现方式，第一衰退倾向因子如下计算：
Figure PCTCN2017118773-appb-000001
其中，GF为第一衰退倾向因子，i为第一预定时间周期，j为第二预定时间周期，R i为第一预定时间周期i内产生预定用户行为的平均次数，R j为第二预定时间周期j内产生预定用户行为的平均次数，其中，第一预定时间周期i的时间起点比第二预定时间周期j的早，a为底数，n为次数阈值，取值为大于或等于0的整数。
结合第一方面，本公开在第一方面的第二种实现方式中，根据产生预定用户行为的时间间隔确定第二衰退倾向因子，包括：
根据第一时间间隔和第二时间间隔的比率确定第二衰退倾向因子，第一时间间隔为最后一次产生预定用户行为距当前时间的时间间隔，第二时间间隔为第三预定时间周期内产生预定用户行为的平均时间间隔，第三预定时间周期的终点在最后一次产生预定用户行为的时间之前。
结合第一方面、第一方面的第二种实现方式，第二衰退倾向因子如下计算：
Figure PCTCN2017118773-appb-000002
其中，R表示第一时间间隔，Mi表示第二时间间隔，第三预定时间周期的终点在最后一次产生预定用户行为的时间之前。
结合第一方面，预定用户行为的总衰退倾向值如下计算：
DI＝b×GF+c×GR
其中，DI为预定用户行为的总衰退倾向值，GF为第一衰退倾向因子，GR为第二衰退倾向因子，b为权重值。
结合第一方面、第一方面的第一种实现方式或第一方面的第二种实现方式，本公开在第一方面的第三种实现方式中，方法还包括：
按照预定用户行为的总衰退倾向值的高低输出预定用户行为的总衰退倾向值小于第一预定值的用户。
第二方面，本公开实施例提供了一种确定用户行为衰退倾向的装置，包括：
第一确定模块，被配置为根据产生预定用户行为的次数确定第一衰退倾向因子；
第二确定模块，被配置为根据产生预定用户行为的时间间隔确定第二衰退倾向因子；
第三确定模块，被配置为基于第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值。
结合第二方面，本公开在第二方面的第一种实现方式中，第一确定模块，包括：
第一确定子模块，被配置为根据在预定时间周期内产生预定用户行为的平均次数变化确定第一衰退倾向因子。
结合第二方面、第二方面的第一种实现方式，第一衰退倾向因子如下计算：
Figure PCTCN2017118773-appb-000003
其中，GF为第一衰退倾向因子，i为第一预定时间周期，j为第二预定时间周期，R i为第一预定时间周期i内产生预定用户行为的平均次数，R j为第二预定时间周期j内产生预定用户行为的平均次数，其中，第一预定时间周期i的时间起点比第二预定时间周期j的早，a为底数，n为次数阈值，取值为大于或等于0的整数。
结合第二方面，本公开在第二方面的第二种实现方式中，第二确定模块，包括：
第二确定子模块，被配置为根据第一时间间隔和第二时间间隔的比率确定第二衰退倾向因子，第一时间间隔为最后一次产生预定用户行为距当前时间的时间间隔，第二时间间隔为第三预定时间周期内产 生预定用户行为的平均时间间隔，第三预定时间周期的终点在最后一次产生预定用户行为的时间之前。
结合第二方面、第二方面的第二种实现方式，第二衰退倾向因子如下计算：
Figure PCTCN2017118773-appb-000004
其中，R表示第一时间间隔，Mi表示第二时间间隔，第三预定时间周期的终点在最后一次产生预定用户行为的时间之前。
结合第二方面，预定用户行为的总衰退倾向值如下计算：
DI＝b×GF+c×GR
其中，DI为预定用户行为的总衰退倾向值，GF为第一衰退倾向因子，GR为第二衰退倾向因子，b为权重值。
结合第二方面、第二方面的第一种实现方式或第二方面的第二种实现方式，本公开在第二方面的第三种实现方式中，装置还包括：
输出模块，被配置为按照预定用户行为的总衰退倾向值的高低输出预定用户行为的总衰退倾向值小于第一预定值的用户。
功能可以通过硬件实现，也可以通过硬件执行相应的软件实现。硬件或软件包括一个或多个与上述功能相对应的模块。
在一个可能的设计中，确定用户行为衰退倾向的装置的结构中包括存储器和处理器，存储器用于存储一条或多条支持确定预定用户行为衰退倾向的装置执行上述第一方面中确定定用户行为衰退倾向的方法的计算机指令，处理器被配置为用于执行存储器中存储的计算机指令。确定用户行为衰退倾向的装置还可以包括通信接口，用于确定用户行为衰退倾向的装置与其他设备或通信网络通信。
第三方面，本公开实施例还提供了一种电子设备，电子设备包括存储器和处理器；其中，存储器用于存储一条或多条计算机指令，其中，一条或多条计算机指令被处理器执行以实现第一方面的确定用户行为衰退倾向的方法。
第三方面，本公开实施例提供了一种计算机可读存储介质，用于存储确定用户行为衰退倾向的装置所用的计算机指令，其包含用于执行上述第一方面中确定用户行为衰退倾向的方法所涉及的计算机指令。
本公开实施例提供的技术方案可以包括以下有益效果：
本公开实施例根据产生预定用户行为的次数确定第一衰退倾向因子，并根据产生预定用户行为的时间间隔确定第二衰退倾向因子，并综合考虑第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值。本公开实施例通过第一衰退因子将用户产生预定用户行为的频率下降作为用户衰退程度刻画的参考指标，能够尽快获取到系统平台中处于衰退期的用户，又进一步通过第二衰退因子将用户近期未产生预定用户行为的时间间隔相对于之前产生预定用户行为的时间间隔的比率作为用户衰退程度刻画的另一参考指标，相较于已有技术中单一的根据用户近期未产生预定行为的时间间隔来说更加全面，并且通过综合以上两种因子，能够更加全面准确的确定用户行为的衰退倾向，从而采取及时干预措施，能大大减少应用平台的损失。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本公开。
附图说明
结合附图，通过以下非限制性实施方式的详细描述，本公开的其它特征、目的和优点将变得更加明显。在附图中：
图1示出根据本公开一实施方式的确定用户行为衰退倾向的方法的流程图；
图2示出根据本公开一实施方式的确定用户行为衰退倾向装置的结构框图；
图3是适于用来实现根据本公开一实施方式的确定用户行为衰退倾向方法的电子设备的结构示意图。
具体实施方式
下文中，将参考附图详细描述本公开的示例性实施方式，以使本领域技术人员可容易地实现它们。此外，为了清楚起见，在附图中省略了与描述示例性实施方式无关的部分。
在本公开中，应理解，诸如“包括”或“具有”等的术语旨在指示本说明书中所公开的特征、数字、步骤、行为、部件、部分或其组合的存在，并且不欲排除一个或多个其他特征、数字、步骤、行为、部件、部分或其组合存在或被添加的可能性。
另外还需要说明的是，在不冲突的情况下，本公开中的实施例及实施例中的特征可以相互组合。下面将参考附图并结合实施例来详细说明本公开。
已有技术对于平台用户衰退倾向的刻画往往只考虑最近产生预定用户行为的时间这一个参量，没有对用户进行针对性判别，衰退倾向的刻画采用的特征非常单一，对衰退倾向的刻画不够准确，比如对于原先产生预定用户行为的时间间隔不同的用户，当他们的未产生预定用户行为的时间间隔相同时，采用已有技术得到的两者的衰退倾向值却相同，这显然不客观。
本公开实施例，根据产生预定用户行为的次数确定第一衰退倾向因子，并根据产生预定用户行为的时间间隔确定第二衰退倾向因子，并综合考虑第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值。通过第一衰退因子将用户产生预定用户行为的频率下降作为用户衰退程度刻画的参考指标，能够尽快获取到系统平台中处于衰退期的用户，又进一步通过第二衰退因子将用户近期未产生预定用户行为的时间间隔相对于之前产生预定用户行为的时间间隔的比率作为用户衰退程度刻画的另一参考指标，相较于已有技术中单一的根据用户近期未产生预定行为的时间间隔来说更加全面，并且通过综合以上两种因子，能够更加全面准确的确定用户行为的衰退倾向，从而采取及时干预措施，能大大减少应用平台的损失。
图1示出根据本公开一实施方式的确定用户行为衰退倾向的方法的流程图。如图1所示，确定用户行为衰退倾向的方法包括以下步 骤S101-S103：
在步骤S101中，根据产生预定用户行为的次数确定第一衰退倾向因子；
在步骤S102中，根据产生预定用户行为的时间间隔确定第二衰退倾向因子；
在步骤S103中，基于第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值。
用户行为可以是用户在系统中的有效行为，包括用户对于系统对象的各种操作，例如，对于商业运营应用平台，用户利用该应用平台进行下单的用户行为。预定用户行为是应用平台欲通过其衡量用户衰退倾向的预定义的用户行为，例如，对于外卖应用平台，衡量用户衰退倾向最好的用户行为可能是用户的下单行为，而对于某产品推销网站应用平台，横向用户衰退倾向最好的用户行为可能是用户的浏览行为等。预定用户行为根据应用平台所涉及内容的不同而可能不同，具体根据实际情况设定。用户在某个时间段内产生预定用户行为的次数减少的时期可以称为衰退期，而用户的预定用户行为目前衰退的程度称为衰退倾向。例如，对于商业运营应用平台，用户在某个时间段内下单的次数减少，则这段时间为用户下单的衰退期，而用户下单次数减少的程度为用户下单的衰退倾向。
本公开实施例中，通过综合考虑用户产生预定用户行为的次数以及时间间隔来确定用户的某一预定用户行为的总衰退倾向值。例如，对于商业运营应用平台，用户在某一个时间段内下单的次数减少，且下单的时间间隔也较长的情况下，可以通过用户下单的总衰退倾向值判定用户在这段时间有衰退倾向，商业运营平台在及时捕捉到用户衰退倾向的情况下，通过各种措施进行干预，防止用户流失。
可以理解的是，上述确定用户衰退倾向的方法不仅仅适用于商业运营应用平台，还可适用于各种供用户使用的应用平台，应用平台可以通过确定用户衰退倾向来反馈用户对本应用平台涉及产品或服务的关注度以及使用情况，进而根据这些反馈信息进一步提高应用平台为用户提供的产品或服务的质量等，能够促进应用平台的进一步发展， 提高用户的体验。
在本实施例的一个可选实现方式中，步骤S101，即根据产生预定用户行为的次数确定第一衰退倾向因子的步骤，进一步包括以下步骤：根据在预定时间周期内产生预定用户行为的平均次数变化确定第一衰退倾向因子。该可选实现方式中，可以通过用户在一段时间内产生预定用户行为的次数变化来确定第一衰退因子，按照时间先后，如果次数逐渐减小，可以推断出用户产生的预定用户行为在减少，而这能够反映出用户在一定程度上呈现出了衰退倾向。例如，用户在一商业运营应用平台上，前一周7天的下单总次数为Ri，而最近一周7天的下单总次数为Rj，且Ri>Rj，此时可以初步确定用户下单的行为处于衰退期，用户有衰退倾向，而衰退倾向的大小与Ri与Rj的差值的绝对值有关。如果Ri>Rj，且Ri与Rj的差值的绝对值较大，说明用户的衰退倾向较大，差值的绝对值较小，说明用户的衰退倾向较小。本实施例通过将用户产生的预定用户行为的次数变化作为衡量用户行为衰退倾向的一个因子，能尽快获取到衰退的平台高价值用户，从而可以根据用户行为的衰退倾向采取及时干预措施，能大大减少应用平台的损失。
在本实施例的一个可选实现方式中，第一衰退倾向因子可如下计算：
Figure PCTCN2017118773-appb-000005
其中，GF为第一衰退倾向因子，i为第一预定时间周期，j为第二预定时间周期，R i为第一预定时间周期i内产生预定用户行为的平均次数，R j为第二预定时间周期j内产生预定用户行为的平均次数，其中，第一预定时间周期i的时间起点比第二预定时间周期j的早，a为底数，n为次数阈值，取值为大于或等于0的整数。
该可选实现方式中，第一预定时间周期i的时间起点比第二预定时间周期j的早，第一预定时间周期i的时间终点可以比第二预定时 间周期j早，也可以相同。例如，第二预定时间周期j可以是前N(N>＝1)天，而第一预定时间周期i为前M(M>＝1)天，M>N；或者第一预定时间周期i为最近1周，而第二预定时间周期j为上一周等，第一预定时间周期i和第一预定时间周期j可根据实际情况进行选择，只要能够使得R i和R j表现出从前往后用户产生预定用户行为的平均次数的变化即可。例如，对于商业运营应用平台，第二预定时间周期j为最近一周，R j则为最近一周7天内的平均下单次数，第一预定时间周期i为最近一个月，R i则为最近一个月内的平均下单次数。该可选实现方式中采用指数函数量化用户产生预定用户行为的次数的下降趋势，在第一预定时间周期内产生预定用户行为的平均次数大于第二预定时间周期内产生预定用户行为的平均次数与次数阈值n之和时，即近期产生预定用户行为的平均次数相较于前期减少了，且减少量大于次数阈值n时，采用底数为a的指数函数量化产生预定用户行为的下降趋势，而第一预定时间周期内产生预定用户行为的平均次数小于或等于第二预定时间周期内产生预定用户行为的平均次数与次数阈值n之和时，即近期产生预定用户行为的平均次数相较于前期而言并未减少，或者减少量依然在次数阈值n范围之内时，则直接置第一衰退倾向因子为0。
该可选实现方式中，i、j、a、n等参数的取值可根据实际情况进行设置。例如，对于外卖类的商业运营应用平台，把连续40天未下单的用户基本认定为流失，所以衰退倾向只刻画未下单时间不超过40天的用户，因此可以根据以下原则确定上述参数的选择：
因为考虑到用户40天未下单可以认定为已经流失，所以可以初步选择i为最近2周的时间周期，而j为最近6周的时间周期。指数函数底数a的确定可以遵循以下原则：第一衰退倾向因子侧重于单量的下降，假如认为用户7日平均下单量下降大于2时，由于单量下降巨大的用户其价值高，因此总衰退倾向值可以以第一衰退倾向因子的值为主导。而为了在用户7日平均下单量下降大于2时，使第一衰退倾向因子的值起主导作用，因此可以考虑使得第一衰退倾向因子的最小值大于或等于第二衰退倾向因子的最大值，这样通过使得a 2约等 于第二衰退倾向因子的最大值，进而确定出底数a的值，当然也还可以根据其他情况对底数a的值进行调整设置，对此本实施例不做任何限制。而次数阈值n可以取大于等于0的整数，具体可视实际情况而定，而且还可以在不同时期调整成不同的值，例如确定用户下单行为的衰退倾向时，促销淡季可以将n的值设置成较大值，而在促销旺季将n的值设置成较小值，因为促销淡季用户下单的次数比促销旺季下单次数少是正常的，因此可以通过对次数阈值n的调整能够更加客观地刻画用户衰退倾向。
该可选实现方式中第一衰退倾向因子的计算公式采用指数函数，使得随着用户产生预定用户行为的次数下降越大，第一衰退倾向因子的值越大，使得衰减倾向增长越快。因为用户产生预定用户行为次数高的用户通常是应用平台的高价值用户，这类用户的衰退倾向需要更及时干预，而通过该可选实现方式中第一衰退因子可以及时监测到用户产生预定用户行为的下降区域并予以及时干预。
在本实施例的另一个可选实现方式中，在步骤102中，即根据产生预定用户行为的时间间隔确定第二衰退倾向因子的步骤，进一步包括以下步骤：根据第一时间间隔和第二时间间隔的比率确定第二衰退倾向因子，第一时间间隔为最后一次产生预定用户行为距当前时间的时间间隔，第二时间间隔为第三预定时间周期内产生预定用户行为的平均时间间隔，第三预定时间周期的终点在最后一次产生预定用户行为的时间之前。该可选实现方式中，通过最近产生预定用户行为距当前时间的时间间隔与之前产生预定用户行为的平均时间间隔的比率来确定第二衰退倾向因子，如果比率越大，则第二衰退倾向因子越大，说明用户的衰退倾向也越大。第一时间间隔可以通过计算当前时间与用户最后一次产生预定行为的时间差值确定，计量单位可以是天、小时等。例如，系统平台记录用户最后产生预定行为的时间为4天前，则距当前时间的时间间隔为4天，而4天前的一段时间(如7天或半个月等)内产生预定用户行为的平均时间间隔为2天，两者的比率为2，这时第二衰退倾向因子的大小由比率2来确定。
在本实施例的一可选实现方式中，第二衰退倾向因子如下计算：
Figure PCTCN2017118773-appb-000006
其中，R表示第一时间间隔，即最后一次产生预定用户行为距当前时间的时间间隔，Mi表示第二时间间隔，即在第三预定时间周期内产生预定用户行为的平均时间间隔，m为时间间隔阈值。
该可选实现方式中，最后一次产生预定用户行为距当前时间的第一时间间隔小于或等于第三预定时间周期内产生预定用户行为的平均时间间隔即第二时间间隔与时间间隔阈值之和时，可以确定用户产生预定用户行为的时间间隔并未增大，或者时间间隔的增大量依然在时间间隔阈值m范围之内时，此时用户处于正常时期，并未进入衰退期，因此可将第二衰退倾向因子置为0；而在最后一次产生预定用户行为距当前时间的第一时间间隔大于第三预定时间周期内产生预定用户行为的平均时间间隔即第二时间间隔与时间间隔阈值之和时，可以确定用户产生预定用户行为的时间间隔增大，并且超出了预先设置的时间间隔阈值m范围之外时，此时用户处于衰退期，因此通过上述对数函数确定第二衰退倾向因子。
第三预定时间周期、m等参数的值可根据实际情况进行设置。假如，对于外卖类的商业运营应用平台，把连续40天未下单的用户基本认定为流失，那么可以将第三预定时间周期设置成40，Mi为用户最后一次下单之前的40天内的平均下单时间间隔。而时间间隔阈值m可以取大于等于0的整数，具体可视实际情况而定，而且还可以在不同时期调整成不同的值，例如在确定用户下单行为的衰退倾向时，促销淡季可以将m的值设置成较大值，而在促销旺季将m的值设置成较小值，因为促销淡季用户下单的时间间隔比促销旺季下单的时间间隔长是正常的，因此可以通过对时间间隔阈值m的调整能够更加客观地刻画用户衰退倾向。
在本实施例的另一可选实现方式中，预定用户行为的总衰退倾向 值如下计算：
DI＝b×GF+c×GR
其中，DI为预定用户行为的总衰退倾向值，GF为第一衰退倾向因子，GR为第二衰退倾向因子，b、c为权重值。
该可选实现方式中，通过将第一衰退倾向因子和第二衰退倾向因子加权后相加来确定总衰退倾向值。b、c的值可以根据实际情况进行设置，对于不同行业、不同情况下第一衰退倾向因子和第二衰退倾向因子的影响力可以不同，有的行业第一衰退倾向因子更能体现用户的衰退倾向，有的行业第二衰退倾向因子可能更能体现用户的衰退倾向，同一行业在不同时期第一衰退倾向因子和第二衰退倾向因子在用户的衰退倾向上的比重也有可能不同，具体需要根据实际情况来判定，并设置参数b、c的值。
在本实施例的另一种可选实现方式中，方法还包括以下步骤：按照预定用户行为的总衰退倾向值的高低输出预定用户行为的总衰退倾向值小于第一预定值的用户。该可选实现方式中，确定出用户的总衰退倾向值以后，如果总衰退倾向值小于第一预定值(可根据实际情况设置)，则说明该用户处于衰退期，可以采取相应的干预措施，以防止用户的流失。按照顺序将总衰退倾向值小于第一预定值的所有用户进行输出，并针对不同的用户或用户群体采用不同的干预策略，可以防止应用平台用户的流失。
例如，对于商业运营应于平台，在选定计算第一衰退因子、第二衰退因子和总衰退倾向值相关的参数并计算出用户的总衰退倾向值后，对于总衰退倾向值高于第一预定值的用户进行营销干预。为了评测这些用户是否会在一定周期内恢复到原先状态，可以采用六周次数未恢复用户率k作为评测指标，其具体计算公式如下：
Figure PCTCN2017118773-appb-000007
下面是针对某商业运营应用平台进行测试得出的数据，下面三张表分别为xx年m月1号，m+1月1号，和m+2月1号得到的总衰退 倾向值排名百分比数据以及相应的k值数据，具体如下：
表1：xx年m月1日
用户排名前百分比 用户数 k值
1％ 95937 89％
2％ 191874 88％
5％ 479687 85％
10％ 959374 81％
20％ 1918748 78％
50％ 4796871 73％
100％ 9593743 67％
表2：xx年m+1月1日
用户排名前百分比 用户数 k值
1％ 82592 88％
2％ 165185 86％
5％ 412963 81％
10％ 825926 75％
20％ 1651853 75％
50％ 4129632 69％
100％ 8259265 60％
表3：xx年m+2月1日
用户排名前百分比 用户数 k值
1％ 109323 81％
2％ 218646 80％
5％ 546616 80％
10％ 1093232 81％
20％ 2186465 78％
50％ 5466164 78％
100％ 10932329 73％
表1-表3中第一列为用户下单行为的总衰倾向值按大小排名后的百分比数据，第二列为对应百分比下的用户数目，第三列为这些用户数的k值。退根据上述数据分析得出，随着用户下单行为的总衰倾向值的百分比数的降低，k值会降低，而排名前2％的用户k值总体大于88％，用户体量约18万，因此可以确定这批用户属于下单次数高，对这批用户进行营销干预比较合适。因此，通过本公开上述实施例得到的用户行为的总衰退倾向值，确定出需要进行干预的用户，并采取 对应的营销干预措施，能够防止应用平台用户的流失，促进应用平台业务的发展。
下述为本公开装置实施例，可以用于执行本公开方法实施例。
图2示出根据本公开一实施方式的确定用户行为衰退倾向的装置的结构框图，该装置可以通过软件、硬件或者两者的结合实现成为电子设备的部分或者全部。如图2所示，确定用户行为衰退倾向的装置包括第一确定模块201、第二确定模块202和第三确定模块203：
第一确定模块201，被配置为根据产生预定用户行为的次数确定第一衰退倾向因子；
第二确定模块202，被配置为根据产生预定用户行为的时间间隔确定第二衰退倾向因子；
第三确定模块203，被配置为基于第一衰退倾向因子和第二衰退倾向因子确定预定用户行为的总衰退倾向值。
用户行为可以是用户在系统中的有效行为，包括用户对于系统对象的各种操作，例如，对于商业运营应用平台，用户利用该应用平台进行下单的用户行为。预定用户行为是应用平台欲通过其衡量用户衰退倾向的预定义的用户行为，例如，对于外卖应用平台，衡量用户衰退倾向最好的用户行为可能是用户的下单行为，而对于某产品推销网站应用平台，横向用户衰退倾向最好的用户行为可能是用户的浏览行为等。预定用户行为根据应用平台所涉及内容的不同而可能不同，具体根据实际情况设定。用户在某个时间段内产生预定用户行为的次数减少的时期可以称为衰退期，而用户的预定用户行为目前衰退的程度称为衰退倾向。例如，对于商业运营应用平台，用户在某个时间段内下单的次数减少，则这段时间为用户下单的衰退期，而用户下单次数减少的程度为用户下单的衰退倾向。
本公开实施例中，通过综合考虑用户产生预定用户行为的次数以及时间间隔来确定用户的某一预定用户行为的总衰退倾向值。例如，对于商业运营应用平台，用户在某一个时间段内下单的次数减少，且下单的时间间隔也较长的情况下，可以通过用户下单的总衰退倾向值判定用户在这段时间有衰退倾向，商业运营平台在及时捕捉到用户衰 退倾向的情况下，通过各种措施进行干预，防止用户流失。
可以理解的是，上述确定用户衰退倾向的方法不仅仅适用于商业运营应用平台，还可适用于各种供用户使用的应用平台，应用平台可以通过确定用户衰退倾向来反馈用户对本应用平台涉及产品或服务的关注度以及使用情况，进而根据这些反馈信息进一步提高应用平台为用户提供的产品或服务的质量等，能够促进应用平台的进一步发展，提高用户的体验。
在本实施例的一个可选实现方式中，第一确定模块201包括：第一确定子模块，被配置为根据在预定时间周期内产生预定用户行为的平均次数变化确定第一衰退倾向因子。该可选实现方式中，可以通过用户在一段时间内产生预定用户行为的次数变化来确定第一衰退因子，按照时间先后，如果次数逐渐减小，可以推断出用户产生的预定用户行为在减少，而这能够反映出用户在一定程度上呈现出了衰退倾向。例如，用户在一商业运营应用平台上，前一周7天的下单总次数为Ri，而最近一周7天的下单总次数为Rj，且Ri>Rj，此时可以初步确定用户下单的行为处于衰退期，用户有衰退倾向，而衰退倾向的大小与Ri与Rj的差值的绝对值有关。如果Ri>Rj，且Ri与Rj的差值的绝对值较大，说明用户的衰退倾向较大，差值的绝对值较小，说明用户的衰退倾向较小。本实施例通过将用户产生的预定用户行为的次数变化作为衡量用户行为衰退倾向的一个因子，能尽快获取到衰退的平台高价值用户，从而可以根据用户行为的衰退倾向采取及时干预措施，能大大减少应用平台的损失。
在本实施例的一个可选实现方式中，第一衰退倾向因子可如下计算：
Figure PCTCN2017118773-appb-000008
其中，GF为第一衰退倾向因子，i为第一预定时间周期，j为第二预定时间周期，R i为第一预定时间周期i内产生预定用户行为的平 均次数，R j为第二预定时间周期j内产生预定用户行为的平均次数，其中，第一预定时间周期i的时间起点比第二预定时间周期j的早，a为底数，n为次数阈值，取值为大于或等于0的整数。
该可选实现方式中，第一预定时间周期i的时间起点比第二预定时间周期j的早，第一预定时间周期i的时间终点可以比第二预定时间周期j早，也可以相同。例如，第二预定时间周期j可以是前N(N>＝1)天，而第一预定时间周期i为前M(M>＝1)天，M>N；或者第一预定时间周期i为最近1周，而第二预定时间周期j为上一周等，第一预定时间周期i和第一预定时间周期j可根据实际情况进行选择，只要能够使得R i和R j表现出从前往后用户产生预定用户行为的平均次数的变化即可。例如，对于商业运营应用平台，第二预定时间周期j为最近一周，R j则为最近一周7天内的平均下单次数，第一预定时间周期i为最近一个月，R i则为最近一个月内的平均下单次数。该可选实现方式中采用指数函数量化用户产生预定用户行为的次数的下降趋势，在第一预定时间周期内产生预定用户行为的平均次数大于第二预定时间周期内产生预定用户行为的平均次数与次数阈值n之和时，即近期产生预定用户行为的平均次数相较于前期减少了，且减少量大于次数阈值n时，采用底数为a的指数函数量化产生预定用户行为的下降趋势，而第一预定时间周期内产生预定用户行为的平均次数小于或等于第二预定时间周期内产生预定用户行为的平均次数与次数阈值n之和时，即近期产生预定用户行为的平均次数相较于前期而言并未减少，或者减少量依然在次数阈值n范围之内时，则直接置第一衰退倾向因子为0。
该可选实现方式中，i、j、a、n等参数的取值可根据实际情况进行设置。例如，对于外卖类的商业运营应用平台，把连续40天未下单的用户基本认定为流失，所以衰退倾向只刻画未下单时间不超过40天的用户，因此可以根据以下原则确定上述参数的选择：
因为考虑到用户40天未下单可以认定为已经流失，所以可以初步选择i为最近2周的时间周期，而j为最近6周的时间周期。指数函数底数a的确定可以遵循以下原则：第一衰退倾向因子侧重于单量 的下降，假如认为用户7日平均下单量下降大于2时，由于单量下降巨大的用户其价值高，因此总衰退倾向值可以以第一衰退倾向因子的值为主导。而为了在用户7日平均下单量下降大于2时，使第一衰退倾向因子的值起主导作用，因此可以考虑使得第一衰退倾向因子的最小值大于或等于第二衰退倾向因子的最大值，这样通过使得a 2约等于第二衰退倾向因子的最大值，进而确定出底数a的值，当然也还可以根据其他情况对底数a的值进行调整设置，对此本实施例不做任何限制。而次数阈值n可以取大于等于0的整数，具体可视实际情况而定，而且还可以在不同时期调整成不同的值，例如确定用户下单行为的衰退倾向时，促销淡季可以将n的值设置成较大值，而在促销旺季将n的值设置成较小值，因为促销淡季用户下单的次数比促销旺季下单次数少是正常的，因此可以通过对次数阈值n的调整能够更加客观地刻画用户衰退倾向。
该可选实现方式中第一衰退倾向因子的计算公式采用指数函数，使得随着用户产生预定用户行为的次数下降越大，第一衰退倾向因子的值越大，使得衰减倾向增长越快。因为用户产生预定用户行为次数高的用户通常是应用平台的高价值用户，这类用户的衰退倾向需要更及时干预，而通过该可选实现方式中第一衰退因子可以及时监测到用户产生预定用户行为的下降区域并予以及时干预。
在本实施例的另一个可选实现方式中，第二确定模块202包括：第二确定子模块，根据第一时间间隔和第二时间间隔的比率确定第二衰退倾向因子，第一时间间隔为最后一次产生预定用户行为距当前时间的时间间隔，第二时间间隔为第三预定时间周期内产生预定用户行为的平均时间间隔，第三预定时间周期的终点在最后一次产生预定用户行为的时间之前。该可选实现方式中，通过最近产生预定用户行为距当前时间的时间间隔与之前产生预定用户行为的平均时间间隔的比率来确定第二衰退倾向因子，如果比率越大，则第二衰退倾向因子越大，说明用户的衰退倾向也越大。第一时间间隔可以通过计算当前时间与用户最后一次产生预定行为的时间差值确定，计量单位可以是天、小时等。例如，系统平台记录用户最后产生预定行为的时间为4 天前，则距当前时间的时间间隔为4天，而4天前的一段时间(如7天或半个月等)内产生预定用户行为的平均时间间隔为2天，两者的比率为2，这时第二衰退倾向因子的大小由比率2来确定。
在本实施例的一可选实现方式中，第二衰退倾向因子如下计算：
Figure PCTCN2017118773-appb-000009
其中，R表示第一时间间隔，即最后一次产生预定用户行为距当前时间的时间间隔，Mi表示第二时间间隔，即在第三预定时间周期内产生预定用户行为的平均时间间隔，m为时间间隔阈值。
该可选实现方式中，最后一次产生预定用户行为距当前时间的第一时间间隔小于或等于第三预定时间周期内产生预定用户行为的平均时间间隔即第二时间间隔与时间间隔阈值之和时，可以确定用户产生预定用户行为的时间间隔并未增大，或者时间间隔的增大量依然在时间间隔阈值m范围之内时，此时用户处于正常时期，并未进入衰退期，因此可将第二衰退倾向因子置为0；而在最后一次产生预定用户行为距当前时间的第一时间间隔大于第三预定时间周期内产生预定用户行为的平均时间间隔第二时间间隔与时间间隔阈值之和时，可以确定用户产生预定用户行为的时间间隔增大，并且超出了预先设置的时间间隔阈值m范围之外时，此时用户处于衰退期，因此通过上述对数函数确定第二衰退倾向因子。
第三预定时间周期、m等参数的值可根据实际情况进行设置。假如，对于外卖类的商业运营应用平台，把连续40天未下单的用户基本认定为流失，那么可以将第三预定时间周期设置成40，Mi为用户最后一次下单之前的40天内的平均下单时间间隔。而时间间隔阈值m可以取大于等于0的整数，具体可视实际情况而定，而且还可以在不同时期调整成不同的值，例如在确定用户下单行为的衰退倾向时，促销淡季可以将m的值设置成较大值，而在促销旺季将m的值设置成 较小值，因为促销淡季用户下单的时间间隔比促销旺季下单的时间间隔长是正常的，因此可以通过对时间间隔阈值m的调整能够更加客观地刻画用户衰退倾向。
在本实施例的另一可选实现方式中，预定用户行为的总衰退倾向值如下计算：
DI＝b×GF+c×GR
其中，DI为预定用户行为的总衰退倾向值，GF为第一衰退倾向因子，GR为第二衰退倾向因子，b、c为权重值。
该可选实现方式中，通过将第一衰退倾向因子和第二衰退倾向因子加权后相加来确定总衰退倾向值。b、c的值可以根据实际情况进行设置，对于不同行业、不同情况下第一衰退倾向因子和第二衰退倾向因子的影响力可以不同，有的行业第一衰退倾向因子更能体现用户的衰退倾向，有的行业第二衰退倾向因子可能更能体现用户的衰退倾向，同一行业在不同时期第一衰退倾向因子和第二衰退倾向因子在用户的衰退倾向上的比重也有可能不同，具体需要根据实际情况来判定，并设置参数b、c的值。
在本实施例的另一种可选实现方式中，装置还包括：输出模块，被配置为按照预定用户行为的总衰退倾向值的高低输出预定用户行为的总衰退倾向值小于第一预定值的用户。该可选实现方式中，确定出用户的总衰退倾向值以后，如果总衰退倾向值小于第一预定值(可根据实际情况设置)，则说明该用户处于衰退期，可以采取相应的干预措施，以防止用户的流失。按照顺序将总衰退倾向值小于第一预定值的所有用户进行输出，并针对不同的用户或用户群体采用不同的干预策略，可以防止应用平台用户的流失。
本领域技术人员可以理解，上述数确定用户行为的衰退倾向的装置还包括一些其他公知结构，例如处理器、存储器等，为了不必要地模糊本公开的实施例，这些公知的结构在图2中未示出。
图3是适于用来实现根据本公开实施方式的确定用户行为的衰退倾向方法的电子设备的结构示意图。
如图3所示，电子设备300包括中央处理单元(CPU)301，其可 以根据存储在只读存储器(ROM)302中的程序或者从存储部分308加载到随机访问存储器(RAM)303中的程序而执行上述图1所示的实施方式中的各种处理。在RAM303中，还存储有系统300操作所需的各种程序和数据。CPU301、ROM302以及RAM303通过总线304彼此相连。输入/输出(I/O)接口305也连接至总线304。
以下部件连接至I/O接口305：包括键盘、鼠标等的输入部分306；包括诸如阴极射线管(CRT)、液晶显示器(LCD)等以及扬声器等的输出部分307；包括硬盘等的存储部分308；以及包括诸如LAN卡、调制解调器等的网络接口卡的通信部分309。通信部分309经由诸如因特网的网络执行通信处理。驱动器310也根据需要连接至I/O接口305。可拆卸介质311，诸如磁盘、光盘、磁光盘、半导体存储器等等，根据需要安装在驱动器310上，以便于从其上读出的计算机程序根据需要被安装入存储部分308。
特别地，根据本公开的实施方式，上文参考图1描述的方法可以被实现为计算机软件程序。例如，本公开的实施方式包括一种计算机程序产品，其包括有形地包含在及其可读介质上的计算机程序，计算机程序包含用于执行图1的键值数据处理方法的程序代码。在这样的实施方式中，该计算机程序可以通过通信部分309从网络上被下载和安装，和/或从可拆卸介质311被安装。
附图中的流程图和框图，图示了按照本公开各种实施方式的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，路程图或框图中的每个方框可以代表一个模块、程序段或代码的一部分，模块、程序段或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序产生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开实施方式中所涉及到的单元或模块可以通过软件的方式实现，也可以通过硬件的方式来实现。所描述的单元或模块也可以设置在处理器中，这些单元或模块的名称在某种情况下并不构成对该单元或模块本身的限定。
作为另一方面，本公开还提供了一种计算机可读存储介质，该计算机可读存储介质可以是上述实施方式中装置中所包含的计算机可读存储介质；也可以是单独存在，未装配入设备中的计算机可读存储介质。计算机可读存储介质存储有一个或者一个以上程序，程序被一个或者一个以上的处理器用来执行描述于本公开的方法。
以上描述仅为本公开的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开中所涉及的发明范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离发明构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。

Claims (16)

  1. 一种确定用户衰退倾向的方法，其中，包括：
    根据产生预定用户行为的次数确定第一衰退倾向因子；
    根据产生所述预定用户行为的时间间隔确定第二衰退倾向因子；
    基于所述第一衰退倾向因子和所述第二衰退倾向因子确定所述预定用户行为的总衰退倾向值。
  2. 根据权利要求1所述的确定用户行为衰退倾向的方法，其中，根据产生预定用户行为的次数确定第一衰退倾向因子，包括：
    根据在预定时间周期内产生所述预定用户行为的平均次数变化确定所述第一衰退倾向因子。
  3. 根据权利要求1-2任一项所述的确定用户行为衰退倾向的方法，其中，所述第一衰退倾向因子如下计算：
    Figure PCTCN2017118773-appb-100001
    其中，GF为所述第一衰退倾向因子，i为第一预定时间周期，j为第二预定时间周期，R i为所述第一预定时间周期i内产生所述预定用户行为的平均次数，R j为所述第二预定时间周期j内产生所述预定用户行为的平均次数，其中，所述第一预定时间周期i的时间起点比所述第二预定时间周期j的早，a为底数，n为次数阈值，取值为大于或等于0的整数。
  4. 根据权利要求1所述的确定用户行为衰退倾向的方法，其中，根据产生所述预定用户行为的时间间隔确定第二衰退倾向因子，包括：
    根据第一时间间隔和第二时间间隔的比率确定所述第二衰退倾向因子，所述第一时间间隔为最后一次产生所述预定用户行为距当前时间的时间间隔，第二时间间隔为第三预定时间周期内产生所述预定用户行为的平均时间间隔，所述第三预定时间周期的终点在最后一次产生所述预定用户行为的时间之前。
  5. 根据权利要求1或4所述的确定用户行为衰退倾向的方法， 其中，所述第二衰退倾向因子如下计算：
    Figure PCTCN2017118773-appb-100002
    其中，R表示第一时间间隔，Mi表示第二时间间隔，所述第三预定时间周期的终点在最后一次产生所述预定用户行为的时间之前。
  6. 根据权利要求1所述的确定用户行为衰退倾向的方法，其中，所述预定用户行为的总衰退倾向值如下计算：
    DI＝b×GF+c×GR
    其中，DI为所述预定用户行为的总衰退倾向值，GF为所述第一衰退倾向因子，所述GR为所述第二衰退倾向因子，b、c为权重值。
  7. 根据权利要求1所述的确定用户行为衰退倾向的方法，其中，所述方法还包括：
    按照所述预定用户行为的总衰退倾向值的高低输出所述预定用户行为的总衰退倾向值小于第一预定值的用户。
  8. 一种确定用户行为衰退倾向的装置，其中，包括：
    第一确定模块，被配置为根据产生预定用户行为的次数确定第一衰退倾向因子；
    第二确定模块，被配置为根据产生所述预定用户行为的时间间隔确定第二衰退倾向因子；
    第三确定模块，被配置为基于所述第一衰退倾向因子和所述第二衰退倾向因子确定所述预定用户行为的总衰退倾向值。
  9. 根据权利要求8所述的确定用户行为衰退倾向的装置，其中，所述第一确定模块，包括：
    第一确定子模块，被配置为根据在预定时间周期内产生所述预定用户行为的平均次数变化确定所述第一衰退倾向因子。
  10. 根据权利要求8-9任一项所述的确定用户行为衰退倾向的装置，其中，所述第一衰退倾向因子如下计算：
    Figure PCTCN2017118773-appb-100003
    其中，GF为所述第一衰退倾向因子，i为第一预定时间周期，j为第二预定时间周期，R i为所述第一预定时间周期i内产生所述预定用户行为的平均次数，R j为所述第二预定时间周期j内产生所述预定用户行为的平均次数，其中，所述第一预定时间周期i的时间起点比所述第二预定时间周期j的早，a为底数，n为次数阈值，取值为大于或等于0的整数。
  11. 根据权利要求8所述的确定用户行为衰退倾向的装置，其中，所述第二确定模块，包括：
    第二确定子模块，被配置为根据第一时间间隔和第二时间间隔的比率确定所述第二衰退倾向因子，所述第一时间间隔为最后一次产生所述预定用户行为距当前时间的时间间隔，第二时间间隔为第三预定时间周期内产生所述预定用户行为的平均时间间隔，所述第三预定时间周期的终点在最后一次产生所述预定用户行为的时间之前。
  12. 根据权利要求8或11所述的确定用户行为衰退倾向的装置，其中，所述第二衰退倾向因子如下计算：
    Figure PCTCN2017118773-appb-100004
    其中，R表示第一时间间隔，Mi表示第二时间间隔，所述第三预定时间周期的终点在最后一次产生所述预定用户行为的时间之前。
  13. 根据权利要求8所述的确定用户行为衰退倾向的装置，其中，所述预定用户行为的总衰退倾向值如下计算：
    DI＝b×GF+c×GR
    其中，DI为所述预定用户行为的总衰退倾向值，GF为所述第一衰退倾向因子，所述GR为所述第二衰退倾向因子，b、c为权重值。
  14. 根据权利要求8所述的确定用户行为衰退倾向的装置，其中，所述装置还包括：
    输出模块，被配置为按照所述预定用户行为的总衰退倾向值的高低输出所述预定用户行为的总衰退倾向值小于第一预定值的用户。
  15. 一种电子设备，其中，包括存储器和处理器；其中，
    所述存储器用于存储一条或多条计算机指令，其中，所述一条或多条计算机指令被所述处理器执行以实现权利要求1-7任一项所述的方法。
  16. 一种计算机可读存储介质，其上存储有计算机指令，其中，该计算机指令被处理器执行时实现权利要求1-7任一项所述的方法。
PCT/CN2017/118773 2017-08-17 2017-12-26 确定用户行为衰退倾向的方法、装置及电子设备 WO2019033677A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710708368.7A CN107527238A (zh) 2017-08-17 2017-08-17 确定用户行为衰退倾向的方法、装置及电子设备
CN201710708368.7 2017-08-17

Publications (1)

Publication Number Publication Date
WO2019033677A1 true WO2019033677A1 (zh) 2019-02-21

Family

ID=60681431

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/118773 WO2019033677A1 (zh) 2017-08-17 2017-12-26 确定用户行为衰退倾向的方法、装置及电子设备

Country Status (2)

Country Link
CN (1) CN107527238A (zh)
WO (1) WO2019033677A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108090785B (zh) * 2017-12-07 2021-03-02 北京小度信息科技有限公司 确定用户行为衰退倾向的方法、装置及电子设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101478450A (zh) * 2009-02-12 2009-07-08 阿里巴巴集团控股有限公司 一种网络用户活跃度的监控方法和服务器
CN105005909A (zh) * 2015-06-17 2015-10-28 深圳市腾讯计算机系统有限公司 预测流失用户的方法及装置
CN105447583A (zh) * 2014-07-30 2016-03-30 华为技术有限公司 一种预测用户离网的方法及装置
CN106919995A (zh) * 2015-12-25 2017-07-04 北京国双科技有限公司 一种判断用户群体流失倾向的方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101478450A (zh) * 2009-02-12 2009-07-08 阿里巴巴集团控股有限公司 一种网络用户活跃度的监控方法和服务器
CN105447583A (zh) * 2014-07-30 2016-03-30 华为技术有限公司 一种预测用户离网的方法及装置
CN105005909A (zh) * 2015-06-17 2015-10-28 深圳市腾讯计算机系统有限公司 预测流失用户的方法及装置
CN106919995A (zh) * 2015-12-25 2017-07-04 北京国双科技有限公司 一种判断用户群体流失倾向的方法及装置

Also Published As

Publication number Publication date
CN107527238A (zh) 2017-12-29

Similar Documents

Publication Publication Date Title
CN108880931B (zh) 用于输出信息的方法和装置
US9508095B2 (en) System and method for optimizing the selection of cloud services based on price and performance
US11669420B2 (en) Monitoring performance of computing systems
CN106294497B (zh) 信息推荐方法和装置
CN107301570B (zh) 业务量预测方法、异常业务量检测方法、装置和电子设备
WO2012057809A2 (en) Methods and apparatus for optimizing advertisement allocation
US9954788B2 (en) Bandwidth estimation based on statistical measures
US20140032379A1 (en) On-shelf availability system and method
US20140052395A1 (en) Method and apparatus for estimating the remaining time required to fully charge a battery, and mobile device incorporating thereof
CN114416512A (zh) 测试方法、装置、电子设备和计算机存储介质
US20160191963A1 (en) Method and system for assessing viewing quality of media objects
CN111753239B (zh) 一种资源分发方法、装置、电子设备及存储介质
CN110569190B (zh) 交易压力测试方法和装置、以及电子设备和可读存储介质
CN110866698A (zh) 用于评定服务提供方的服务分值的装置
JP7085486B2 (ja) ユーザ挙動データの処理方法及び装置
US8930773B2 (en) Determining root cause
CN110738508A (zh) 一种数据分析方法和装置
WO2019033677A1 (zh) 确定用户行为衰退倾向的方法、装置及电子设备
CN106649415B (zh) 客户端数据检查更新方法及装置
US8630890B2 (en) Product migration analysis using data mining by applying a time-series mathematical model
JP2019524000A (ja) ビデオのための相互雑音推定
JP4498666B2 (ja) 予測装置、予測プログラム、および記録媒体
JP5873219B2 (ja) 情報処理装置、電子クーポン処理方法およびプログラム
CN114553614B (zh) 带宽成本估算方法、装置、设备、介质和程序产品
CN111666535B (zh) 确定用户活跃时长的方法、装置、电子设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17921468

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 11.09.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 17921468

Country of ref document: EP

Kind code of ref document: A1