CN115983502A

CN115983502A - Data processing method, apparatus and medium

Info

Publication number: CN115983502A
Application number: CN202310262157.0A
Authority: CN
Inventors: 孙谦; 过群; 赵义雪莹; 杨博; 戴小飞
Original assignee: Beijing Xiaomi Mobile Software Co Ltd
Current assignee: Beijing Xiaomi Mobile Software Co Ltd
Priority date: 2023-03-16
Filing date: 2023-03-16
Publication date: 2023-04-18
Anticipated expiration: 2043-03-16
Also published as: CN115983502B

Abstract

The present disclosure relates to a data processing method, apparatus, and medium, the method comprising: acquiring use state data of a target object in a historical period; determining a characteristic data sequence corresponding to the use state data in each characteristic dimension; for each characteristic dimension, determining a predicted characteristic parameter corresponding to the characteristic data sequence under the characteristic dimension according to a parameter prediction model corresponding to the characteristic dimension; and determining demand forecasting data of the target object according to the forecasting characteristic parameters and the number of users of the target object. Therefore, forecasting can be carried out by cutting in from the characteristic dimension of the demand influencing the target object, and the accuracy of the finally determined demand forecasting data of the target object can be ensured. Meanwhile, the scheme can embody the logic of target object replacement and conversion in the prediction process with clear structure, thereby enhancing the interpretability of the prediction process.

Description

Data processing method, apparatus and medium

Technical Field

The present disclosure relates to the field of data processing, and in particular, to a data processing method, apparatus, and medium.

Background

Accurate forecasts of demand are important in supply chain management, and accurate forecasts of demand quantities can facilitate determining production and distribution quantities, etc. of supply objects. Therefore, how to accurately predict the demand of the supply object in the future period is a problem that needs to be solved.

Disclosure of Invention

To overcome the problems in the related art, the present disclosure provides a data processing method, apparatus, and medium.

According to a first aspect of the embodiments of the present disclosure, there is provided a data processing method, including:

acquiring use state data of a target object in a historical period;

determining a characteristic data sequence corresponding to the use state data under each characteristic dimension;

for each characteristic dimension, determining a predicted characteristic parameter corresponding to the characteristic data sequence under the characteristic dimension according to a parameter prediction model corresponding to the characteristic dimension;

and determining demand forecasting data of the target object according to the forecasting characteristic parameters and the number of users of the target object.

Optionally, the target object corresponds to multiple types, and the feature dimensions include an active user feature, an object replacement cycle feature and a type replacement flow direction feature in each type;

the determining demand forecasting data of the target object according to the forecasting characteristic parameters and the number of users of the target object comprises the following steps:

determining an association type of the target type according to the type change flow direction feature under the target type, wherein the target type is any one of the multiple types, and the association type of the target type is used for representing the type before the target object is changed when the type after the target object is changed is the target type; the type change flow direction feature under the target type is used for representing the association relationship between the association type and the target type;

determining the number of active users of the target object in the association type according to the number of users of the target object in each association type and a predicted characteristic parameter corresponding to the characteristic of the active users in the association type, and determining the predicted replacement number of the target object in the association type according to the number of active users of the target object in the association type and a predicted characteristic parameter corresponding to the characteristic of the object replacement cycle in the association type;

and determining demand forecasting data of the target object of the target type according to the forecasting replacement number of the target object of the target type under each association type.

Optionally, the determining, according to the number of active users of the target object in the association type and a predicted feature parameter corresponding to an object replacement cycle feature in the association type, a predicted replacement number of the target object in the association type includes:

determining a ratio of the number of active users of the target object in the association type to a predicted feature parameter corresponding to an object replacement cycle feature in the association type as a predicted replacement number of the target object in the association type.

Optionally, the parameter prediction model comprises a time series prediction submodel, an integrated moving average autoregressive submodel and a fusion submodel;

the determining the predicted characteristic parameters corresponding to the characteristic data sequence under the characteristic dimension according to the parameter prediction model corresponding to the characteristic dimension includes:

determining a trend item parameter and a period item parameter corresponding to the characteristic data sequence under the characteristic dimension based on the time sequence prediction submodel, and determining a first prediction parameter according to the trend item parameter and the period item parameter;

based on the integrated moving average autoregressive submodel, performing fitting calculation on the characteristic data sequence under the characteristic dimension to obtain a second prediction parameter;

and fusing the first prediction parameter and the second prediction parameter based on the fusion sub-model to obtain the prediction characteristic parameter.

Optionally, before the step of determining demand amount prediction data of the target object according to the prediction feature parameter and the number of users of the target object, the method further includes:

receiving an input adjustment parameter under a target feature dimension, wherein the adjustment parameter is used for representing a target average value of a prediction feature parameter under the target feature dimension;

and adjusting the predicted characteristic parameters under the target characteristic dimension according to the adjustment parameters so as to enable the average value of the adjusted predicted characteristic parameters to reach the target average value.

Optionally, the adjusting the predicted feature parameter under the target feature dimension according to the adjustment parameter includes:

determining an average value of the predicted characteristic parameters under the target characteristic dimension, and taking a ratio of the adjustment parameter to the average value as an adjustment proportion;

and multiplying each numerical value in the predicted characteristic parameters under the target characteristic dimension by the adjustment proportion to obtain the adjusted predicted characteristic parameters.

Optionally, the method further comprises:

acquiring corresponding distribution proportions of the target object under a plurality of warehousing platforms;

and determining the demand distribution quantity of the target objects under each warehousing platform according to the demand prediction data of the target objects and the distribution proportion.

Optionally, the feature dimensions comprise active user features; a plurality of target sub-objects are correspondingly arranged under the target object, and each target sub-object is electronic equipment provided with an application program;

the corresponding characteristic data sequence under the active user characteristic is determined by the following method:

determining an operating parameter of each target sub-object in the target object at each sampling moment in the historical time period according to the use state data of the target object, wherein the operating parameter comprises the use frequency of a target application program and/or the screen use duration ratio of the target sub-object;

for each sampling moment, determining the ratio of the number of the target sub-objects of which the operating parameters are greater than a parameter threshold to the total number of the target sub-objects at the sampling moment as the active user characteristics at the sampling moment;

and taking a sequence obtained by arranging the active user characteristics at each sampling moment according to the sampling moments as a characteristic data sequence corresponding to the active user characteristics.

According to a second aspect of the embodiments of the present disclosure, there is provided a data processing apparatus including:

the device comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is configured to acquire the use state data of a target object in a historical period;

a first determining module configured to determine a feature data sequence corresponding to the usage state data in each feature dimension;

a second determining module, configured to determine, for each of the feature dimensions, a predicted feature parameter corresponding to the feature data sequence in the feature dimension according to a parameter prediction model corresponding to the feature dimension;

a third determination module configured to determine demand prediction data of the target object according to the prediction feature parameter and the number of users of the target object.

the third determining module includes:

a first determining sub-module, configured to determine an association type of a target type according to the type change flow direction feature under the target type, where the target type is any one of the multiple types, and the association type of the target type is used to indicate a type before the target object is changed when a type after the target object is changed is the target type; the type change flow direction feature under the target type is used for representing the association relationship between the association type and the target type;

a second determining sub-module, configured to determine, according to the number of users of the target object in each of the association types and a predicted feature parameter corresponding to an active user feature in the association type, the number of active users of the target object in the association type, and determine, according to the number of active users of the target object in the association type and a predicted feature parameter corresponding to an object replacement cycle feature in the association type, a predicted replacement number of the target object in the association type;

a third determining sub-module configured to determine demand prediction data of the target object of the target type according to the predicted replacement number of the target object of the target type in each of the association types.

Optionally, the second determination submodule is further configured to:

the second determining module includes:

the first processing submodule is configured to determine a trend item parameter and a period item parameter corresponding to the characteristic data sequence under the characteristic dimension based on the time series prediction submodel, and determine a first prediction parameter according to the trend item parameter and the period item parameter;

a second processing submodule configured to perform fitting calculation on the feature data sequence under the feature dimension based on the integrated moving average autoregressive submodel to obtain a second prediction parameter;

a fusion submodule configured to fuse the first prediction parameter and the second prediction parameter based on the fusion submodel to obtain the prediction feature parameter.

Optionally, the apparatus further comprises:

a receiving module configured to receive an input adjustment parameter in a target feature dimension before the third determining module determines the demand amount prediction data of the target object according to the prediction feature parameter and the number of users of the target object, wherein the adjustment parameter is used for representing a target average value of the prediction feature parameter in the target feature dimension;

and the adjusting module is configured to adjust the predicted characteristic parameters under the target characteristic dimension according to the adjusting parameters so that the average value of the adjusted predicted characteristic parameters reaches the target average value.

Optionally, the adjusting module includes:

a fourth determining submodule configured to determine an average value of the predicted feature parameters in the target feature dimension, and use a ratio of the adjustment parameter to the average value as an adjustment ratio;

a fifth determining submodule configured to multiply each value of the predicted feature parameter in the target feature dimension by the adjustment ratio to obtain an adjusted predicted feature parameter.

Optionally, the apparatus further comprises:

the second acquisition module is configured to acquire the corresponding distribution proportion of the target object under a plurality of warehousing platforms;

and the fourth determining module is configured to determine the distribution quantity of the demand of the target object under each warehousing platform according to the demand prediction data of the target object and the distribution proportion.

Optionally, the feature dimension comprises an active user feature; a plurality of target sub-objects are correspondingly arranged under the target object, and each target sub-object is electronic equipment provided with an application program;

for each sampling moment, determining the ratio of the number of the target sub-objects of which the operating parameters are greater than a parameter threshold value to the total number of the target sub-objects at the sampling moment as the active user characteristics at the sampling moment;

According to a third aspect of the embodiments of the present disclosure, there is provided a data processing apparatus including:

a processor;

a memory for storing processor-executable instructions;

wherein the processor is configured to:

acquiring use state data of a target object in a historical period;

for each feature dimension, determining a predicted feature parameter corresponding to the feature data sequence under the feature dimension according to a parameter prediction model corresponding to the feature dimension;

According to a fourth aspect of embodiments of the present disclosure, there is provided a computer-readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the steps of the data processing method of any one of the above first aspects.

The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects:

in the technical scheme, the predicted characteristic parameters of the target object under the characteristic dimension can be determined through the use state data of the target object in the historical period, that is, the characteristic dimension influencing the demand of the target object can be switched in for prediction, and the characteristic dimension is the characteristic clearly related to the demand of the target object, so that the accuracy of the finally determined demand prediction data of the target object can be ensured. Meanwhile, the scheme can embody the logic of target object replacement and conversion in the prediction process with clear structure, thereby enhancing the interpretable capacity of the prediction process, leading the demand forecast data of the target object to be capable of reasoning and explaining based on the prediction process, and further ensuring the accuracy of the demand forecast data of the target object.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure.

FIG. 1 is a flow chart illustrating a method of data processing in accordance with an exemplary embodiment.

FIG. 2 is a schematic illustration of type change flow direction characteristics for various types determined according to an exemplary embodiment.

FIG. 3 is a predictive schematic illustrating a determined predicted number of replacements, according to an exemplary embodiment.

FIG. 4 is a schematic illustration of demand forecast data, according to an exemplary embodiment.

FIG. 5 is a block diagram illustrating a data processing apparatus according to an example embodiment.

FIG. 6 is a block diagram illustrating an apparatus for data processing in accordance with an example embodiment.

FIG. 7 is a block diagram illustrating an apparatus for data processing in accordance with an example embodiment.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

It should be noted that all actions of acquiring signals, information or data in the present application are performed under the premise of complying with the corresponding data protection regulation policy of the country of the location and obtaining the authorization given by the owner of the corresponding device.

With the development of machine learning, in the related art, a sequence formed by demand quantities of supply objects corresponding to different time points is used as an input, and a predicted demand quantity is used as an output to predict the demand quantity on the basis of a machine learning model. However, the process predicted in the above prediction method is usually a black box model, that is, the predicted demand can only be directly determined based on the model by using the demand data in the historical period, the whole process of demand prediction is realized by a machine learning model, and it is difficult to intuitively know the processing process undergone by demand prediction, so that it is difficult to explain the accuracy of the predicted demand and the determination process, that is, it is difficult to explain the prediction result. Based on this, the present disclosure provides the following embodiments to solve the above-described problems.

Fig. 1 is a flow chart illustrating a method of data processing according to an exemplary embodiment, which may include the following steps, as shown in fig. 1.

In step 11, usage status data of the target object within the history period is acquired.

The target object may be a supply object for demand prediction, such as an electronic device like a mobile phone or a tablet. The history period may be set by the user based on the actual application scenario, which is not limited. In this embodiment, the usage status data may be operation data and configuration data of the target object during the usage process, for example, the operation data may be data indicating that the target object operates bright screen and dark screen during the usage process, the number of operations on some application programs, and the like. The configuration data may be a device identification code, a SIM card number, etc.

In step 12, a sequence of feature data corresponding in each feature dimension using the state data is determined. Wherein the feature dimension can be one or more, each feature dimension being used to represent a dimension that has an impact on the demand for the target object. In this embodiment, the use state data may be converted to determine a feature data sequence corresponding to each feature dimension, for example, the feature data sequence may be a time sequence, that is, a sequence in which values in the same feature dimension are arranged according to the occurrence time sequence, and then data in the feature data sequence may represent a change feature of the feature dimension with time.

In step 13, for each feature dimension, according to the parameter prediction model corresponding to the feature dimension, the predicted feature parameter corresponding to the feature data sequence under the feature dimension is determined.

In this embodiment, a parameter prediction model corresponding to each feature dimension may be trained to determine a corresponding predicted feature parameter for the feature dimension in a future time period. Wherein the structure of each parametric prediction model may be the same.

For example, the feature dimension may be an object replacement period feature of the target object, and if the object replacement period is longer, the demand of the target object in the future period is relatively smaller, and if the object replacement period is shorter, the demand of the target object in the future period is relatively larger. In this step, a time series corresponding to the object replacement cycle characteristic corresponding to the target object in the history period may be determined based on the usage state data in the history period, so that the object replacement cycle of the target object in the future period may be predicted based on the time series, thereby further predicting the demand of the target object.

In step 14, demand prediction data of the target object is determined based on the prediction feature parameters and the number of users of the target object.

The user quantity of the target object may represent the current corresponding use quantity of the target object, that is, the stock quantity user of the target object, so that the quantity of the target object to be replaced is determined based on the stock quantity user, and the demand quantity prediction data of the target object is obtained.

Therefore, in the technical scheme, the prediction characteristic parameters of the target object under the characteristic dimension can be determined through the use state data of the target object in the historical period, namely, the prediction can be carried out by switching in from the characteristic dimension which influences the demand of the target object, and the characteristic dimension is the characteristic which is definitely related to the demand of the target object, so that the accuracy of the finally determined demand prediction data of the target object can be ensured. Meanwhile, the scheme can embody the logic of target object replacement and conversion in the prediction process with clear structure, thereby enhancing the interpretable capacity of the prediction process, leading the demand forecast data of the target object to be capable of reasoning and explaining based on the prediction process, and further ensuring the accuracy of the demand forecast data of the target object.

In one possible embodiment, the target object corresponds to a plurality of types, and the feature dimensions comprise an active user feature, an object replacement cycle feature and a type replacement flow direction feature under each type. Wherein the type can be used to represent a brand classification corresponding to the target object. The active user characteristics can be used for representing the proportion of active users in the use users corresponding to the target object, namely the proportion of users which are possible to replace the target object in the use users; the object replacement cycle characteristic can be used for representing the activity of replacing the target object; the type change flow direction feature may be used to indicate a conversion rate when the target objects of the respective types are changed, for example, the type of the target object after the target object of the type a is changed may be the type a or the type B.

Accordingly, an exemplary implementation of the determining demand forecast data of the target object according to the forecast characteristic parameter and the number of users of the target object is as follows, and the step may include:

determining an association type of the target type according to the type change flow direction feature under the target type, wherein the target type is any one of the multiple types, and the association type of the target type is used for indicating the type of the target object before the target object is changed when the type of the target object after the target object is changed is the target type; and the type change flow direction feature under the target type is used for representing the association relationship between the association type and the target type.

When the target object is replaced, as an example, the type before replacement may be used as an outflow type, the type after replacement may be used as an inflow type, and the type replacement flow direction feature may include a conversion rate between a target object of an outflow type and a target object of an inflow type.

For example, the type of the target object may include A, B, C, D, and the target type may be a, where a conversion rate of replacing the type a with the type a in the type replacement flow direction feature corresponding to the type a is m1, a conversion rate of replacing the type B with the type a is m2, and a conversion rate of replacing the type C with the type a is m3, and then for the target type a, the association type of the target type a may be determined to be A, B, C according to the type replacement flow direction feature under the type.

As described above, for each feature dimension, the predicted feature parameter corresponding to the feature data sequence in the feature dimension may be determined according to the parameter prediction model corresponding to the feature dimension. For the type change flow direction feature, a corresponding predicted feature parameter may be determined according to a data feature sequence of the target object under the type change flow direction feature in a history period, where the predicted feature parameter is as shown in fig. 2, and includes the type change flow direction feature under each type, for example, when the outflow type is type a, the predicted conversion rate of the corresponding inflow type is 38.0% for type a, 27.8% for type B, 3.0% for type C, 10.0% for type D, 8.0% for type E, 10.5% for type F, and 2.7% for type G. When the outflow type is type B, the predicted conversion rate corresponding to each inflow type can be determined, which is not described herein again.

Determining the number of active users of the target object under the association type according to the number of users of the target object under each association type and a prediction characteristic parameter corresponding to the active user characteristic under the association type, and determining the predicted replacement number of the target object under the association type according to the number of active users of the target object under the association type and a prediction characteristic parameter corresponding to the object replacement cycle characteristic under the association type.

For each of the association types, the predicted characteristic parameter corresponding to the active user characteristic in the association type is used to represent the proportion of active users in the association type, and then the product of the number of users of the target object in the association type and the predicted characteristic parameter corresponding to the active user characteristic in the association type may be determined as the number of active users of the target object in the association type, that is, the number of users who may replace the target object in the association type. The larger the number of active users of the target object in the association type is, the larger the demand of the target object in the association type is. The number of users of the target object in the association type can determine the type of the device used by each user through the device identification code, so that the number of users corresponding to the device with the type as the association type is used as the number of users of the target object in the association type.

As an example, a ratio of the number of active users of the target object in the association type to a predicted feature parameter corresponding to an object replacement cycle feature in the association type may be determined as a predicted replacement number of the target object in the association type. The larger the predicted replacement number of the target object in the association type is, the larger the required amount of the target object in the association type is.

For example, if the annual prediction is taken as an example, the predicted number of active users in the association type is 6 hundred million, and the predicted characteristic parameter corresponding to the object replacement cycle characteristic is 2 years, it can be estimated that the predicted replacement number of the target object in the association type in the next year is 3 (6/2) million, that is, the number of target objects that are likely to be replaced in the association type in the next year.

It should be noted that the cycle division of the predicted cycle coincides with the cycle division in the target replacement cycle feature, and if the predicted replacement number of the next month is to be predicted, the unit of the predicted feature parameter corresponding to the target replacement cycle may be a month unit. Similarly, it may be default that the unit of the predicted feature parameter corresponding to the target replacement cycle feature is a month unit, the predicted active user number is a for one year of the prediction cycle, the predicted feature parameter corresponding to the target replacement cycle feature is b months, and the predicted replacement number corresponding to the next year may be represented as a/b × 12. For another example, if the prediction period is one quarter, the number of active users predicted is m, and the parameter of the prediction characteristic corresponding to the object replacement period characteristic of the object replacement period characteristic is n months, the predicted replacement number corresponding to the next quarter may be represented as m/n × 3.

As shown in fig. 3, a diagram of a prediction of a predicted replacement number determined according to one embodiment of the present disclosure is shown. For example, if it is currently 2021 year 12 month, the prediction feature parameter corresponding to the currently predicted active user feature of type a is x, and the number of active users is determined to be 25494 based on the product of the number of inventory users of type a obtained in 2021 year 12 month and x, as shown in fig. 3, where the prediction feature parameter corresponding to the object replacement cycle feature of type a predicted based on the scheme of the present disclosure is 46.8, the predicted replacement number corresponding to type a in 2022 year may be 6537. The number of the inventory users under the type A can determine the type of the equipment used by each user through the equipment identification code, so that the number of the users corresponding to the equipment with the type A is used as the number of the inventory users under the type A. In the same way, the predicted replacement number corresponding to each type can be further determined.

And then determining demand forecasting data of the target object of the target type according to the forecasting replacement number of the target object of the target type under each association type.

For each association type, the product of the predicted replacement number of the target object under the association type and the conversion rate of the association type to the target type can be determined as the demand for the target object of the target type generated when the target object of the association type is replaced. Further, the sum of the demand quantities of the target objects of the target types determined under the association types is determined as demand quantity prediction data of the target objects of the target types.

Following the above example, it may be determined that the demand for type a to type a is 2484 (6537 x 38.0%), type B to type a is 1701 (6493 x 26.2%), type C to type a is 898 (5252 x 17.1%), type D to type a is 741 (3567 x 20.8%), type E to type a is 492 (4391 x 11.2%), type F to type a is 336 (2714 x 12.4%), and type G to type a is 144 (795 x 18.2%). Thus, the demand amount prediction data corresponding to the type a can be obtained as 6796. The calculation method for each target type is the same, as shown in fig. 4, and the specific calculation method is not described herein again.

In a possible embodiment, the determination of the demand prediction data may be performed for some types of target objects, or may be performed for each type of target object. In the case where the determination of the demand amount prediction data is performed for each type of target object, the sum of the demand amount prediction data of the target objects of the respective types may also be used as the total demand amount prediction data of the target object. And the demand ratio of the target object of a certain type may also be determined according to the ratio of the demand prediction data of the target object of the certain type to the comprehensive demand prediction data, as shown in fig. 4, so as to perform comprehensive analysis on the target objects of the various types.

Therefore, by the technical scheme, the number of users who possibly replace the target object under the target type can be embodied, the estimated number of the target objects which are replaced under the plurality of users can be further determined, the conversion logic among the target objects of various types can be embodied, the demand of the target object is determined based on the estimated number and the conversion during replacement among the target objects of various types, the demand can be determined through a clear demand evolution process, the interpretability of the demand prediction data of the determined target object is improved, the effectiveness of each step in the demand prediction process is guaranteed, the accuracy of the determined demand prediction data can be guaranteed, and reliable data support is provided for guiding production or distribution of the target object in the follow-up process. Meanwhile, in the process of determining the demand forecast data, forecast analysis can be performed from a plurality of sub-stages in the demand evolution process, and the fitting degree of the demand forecast and the real application scene is improved, so that the usability of the demand forecast data is improved.

In one possible embodiment, the feature dimensions include active user features; a plurality of target sub-objects are correspondingly arranged under the target object, each target sub-object is an electronic device provided with an application program, and for example, each target sub-object can be a mobile phone device;

and determining the operating parameters of each target sub-object under the target object at each sampling moment in the historical period according to the use state data of the target object, wherein the operating parameters comprise the use frequency of a target application program and/or the screen use duration ratio of the target sub-object.

The use state data may include data of operating bright screen and dark screen of each target sub-object during the use process in the historical period, the number of operations on some application programs, and the like. For example, the target application may be set according to an actual application scenario, such as an application that a user uses daily, such as a general social application or a general short video application.

As an example, the history period may be the last year, the sampling time may be the end of each month, the frequency of using the target application of the target sub-object at the sampling time may be the sum of the frequencies of using the target application in the month to which the sampling time belongs, and the ratio of the screen usage duration of the target sub-object at the sampling time may be expressed as the ratio of the screen-on duration of the target sub-object in the month to which the sampling time belongs to the total screen operation duration.

Then, for each sampling moment, determining the ratio of the number of the target sub-objects of which the operation parameters are greater than the parameter threshold to the total number of the target sub-objects at the sampling moment as the active user characteristics at the sampling moment;

In the above example, if the usage frequency of the target application program in the target sub-object is high and the occupation ratio of the screen usage duration is high, it indicates that the target sub-object is an object used by the user daily, that is, the user corresponding to the target sub-object may be considered as an active user, and if the usage frequency of the target application program in the target sub-object is low and the occupation ratio of the screen usage duration is low, it indicates that the target sub-object is not used frequently, for example, the target sub-object may be a standby device of the user.

In this embodiment, it may be considered that a user corresponding to the target sub-object whose operation parameter is greater than the parameter threshold is an active user, and for each sampling time, a ratio of the number of the target sub-objects whose operation parameter is greater than the parameter threshold to the total number of the target sub-objects at the sampling time is determined as an active user characteristic at the sampling time, that is, a proportion of the active user at each sampling time is determined. And then sequencing according to the sampling time to obtain a characteristic data sequence corresponding to the active user characteristic.

If the active user features at the sampling time corresponding to each month in the year are respectively A1-a12, the feature data sequence corresponding to the determined active user features can be represented as { A1, A2, A3, A4, A5, A6, A7, A8, A9, a10, a11, a12}.

Therefore, by the technical scheme, the time sequence corresponding to the characteristic dimension of the active user characteristic can be extracted based on the use state data of the target object obtained in the historical period, so that the characteristic parameter under the characteristic dimension can be predicted based on the time sequence, the relevance between the active user characteristic and the demand of the target object is ensured, and the data processing amount in the characteristic parameter prediction process corresponding to the active user characteristic is reduced.

In a possible embodiment, the feature dimension includes an object replacement cycle feature, and a corresponding feature data sequence under the object replacement cycle feature is determined by:

and determining the corresponding replacement number of each target sub-object in the target object at each sampling moment in the historical period according to the use state data of the target object.

As in the above example, the replacement number corresponding to each sampling time may be the number of target objects replaced in the month to which the sampling time belongs, the number of target objects replaced in the month may be determined based on the usage status data of the target objects, and if the device identification codes corresponding to the same SIM card in the month are different, it indicates that the target sub-object used by the user corresponding to the SIM card is replaced, and at this time, the number of replacement times may be recorded once. In the same manner, the total number of times of replacement of the target object in the month, that is, the number of replacements can be determined.

As another example, in order to determine the replacement number more smoothly, the replacement number may be counted in a time period of a quarter, a year, and the like, which is longer than the time period to which the sampling time belongs, and the replacement number corresponding to the sampling time in the time period may be determined based on an averaging process, where if the replacement number may be counted in a half-year time period to obtain a total number of half-years, the replacement number corresponding to the sampling time per month in the year is the total number of replacement/6 in the half-year, or if the replacement number is counted in a quarter time period to obtain a total number of quarters, the replacement number corresponding to the sampling time per month in the quarter is the total number of quarter/3, and if the replacement number corresponding to the sampling time per month in the year is the total number of year/12. The time period for performing statistics can be limited according to the actual application scenario.

Then, for each sampling moment, determining the ratio of the replacement number to the total number of the target sub-objects corresponding to the active user at the sampling moment as the object replacement cycle characteristic of the sampling moment;

and taking a sequence obtained by arranging the object replacement cycle characteristics at each sampling moment according to the sampling moments as a characteristic data sequence corresponding to the object replacement cycle characteristics.

In the above example, the total number of the target sub-objects corresponding to the active user at the sampling time may be the total number of the target objects corresponding to the active user in the month to which the sampling time belongs, where the identification and determination manner of the active user is described in detail above, and is not described herein again.

The ratio of the replacement number to the total number of the target sub-objects of the active user corresponding to the sampling time is used as the object replacement cycle characteristic of the sampling time, and the replacement activity corresponding to the target object can be represented.

In one possible embodiment, the target object corresponds to a plurality of types, and exemplarily the types can be used to represent brand classifications of different target objects, the feature dimension includes a type change flow direction feature, and a corresponding feature data sequence under the type change flow direction feature is determined by:

and determining the type change flow direction characteristic corresponding to each target sub-object in the target object at each sampling moment in the historical period according to the use state data of the target object.

The method includes the steps that when each target object is determined to be replaced, the types corresponding to the target object before and after replacement are determined, for example, the same SIM card number in data corresponding to the current sampling time is determined to correspond to different device identification numbers a and B according to the use state data, if the time corresponding to the device identification number a is earlier than the time corresponding to the device identification number B, the corresponding types can be determined to be type A and type B respectively according to the device identification numbers a and B, and the determined replacement flow direction for the replacement is type A and type B. Similarly, a change flow direction can be determined for each change in the above manner, and the ratio of the number of times of changing the outflow type to the inflow type in the change flow direction to the number of times of changing the outflow type is used as the conversion rate corresponding to the change flow direction. Such as the conversion rate of type a to type B, i.e. the ratio of the number of times type a is changed to type B to the total number of times target objects of that type a are changed.

And arranging the type change flow direction characteristics of each sampling moment according to the sampling moments to obtain a sequence, and using the sequence as a characteristic data sequence corresponding to the type change flow direction characteristics. Thus, the feature data sequence corresponding to the type change flow direction feature can be represented as a two-dimensional feature, i.e., a matrix of N times N corresponding to each sampling time, where N is the total number of types of the target object, the secondiGo to the firstjMatrix values of the columns representiThe type is outflow type, the secondjOne type is the conversion rate of the change flow direction of the inflow type.

In one possible embodiment, the parametric prediction model includes a time series predictor model, an integrated moving average autoregressive submodel, and a fusion submodel;

an exemplary implementation manner of determining the predicted feature parameter corresponding to the feature data sequence under the feature dimension according to the parameter prediction model corresponding to the feature dimension is as follows, and the step may include:

and determining a trend item parameter and a period item parameter corresponding to the characteristic data sequence under the characteristic dimension based on the time sequence prediction submodel, and determining a first prediction parameter according to the trend item parameter and the period item parameter.

The time series prediction submodel can be realized by an optimized Prophet model representing long term, period and trend. In this embodiment, the first prediction parameter may be determined by:

wherein the content of the first and second substances,

for indicating a corresponding first prediction parameter for time t, <' >>

Representing a prediction error for a model residual term; />

And expressing a trend term parameter for expressing the overall variation trend of the aperiodic, wherein a saturation growth model based on logistic regression is adopted, and the trend term parameter can be determined by the following formula:

wherein the content of the first and second substances,

indicates the time, is>

Represents the bearing capacity (upper value limit) and/or the bearing capacity (upper value limit)>

For indicating a growth rate>

For representing the offset parameter. The calculation mode can be calculated by adopting a determination mode of general trend item parameters in a Prophet model in the field, and details are not repeated herein.

The term parameter representing period can be used to represent the period variation such as month, quarter, year, etc., and can be a period prediction model based on Fourier series, such as can be determined by the following formula:

wherein

The number of days of the cycle is indicated,Nindicates a prediction sensitivity, <' > is>

And &>

For representing coefficients in a Fourier series, the model can be fitted by an L-BFGS algorithm.

And performing fitting calculation on the characteristic data sequence under the characteristic dimension based on the integrated moving average autoregressive submodel to obtain a second prediction parameter.

Wherein the integrated moving average autoregressive submodel may be implemented based on an ARIMA model representing recent trends. It can be determined by:

wherein the content of the first and second substances,

for representing a second prediction parameter, <' > or>

、/>

、/>

Etc. are used to represent the coefficients of the autoregressive term>

、

、/>

Etc. are used to indicate an error term>

、/>

Etc. are used to represent the coefficients of the error term. The d value can be determined from all the training data during the fitting process of the model, and can be used to represent the number of differences that make the sequence stationary, e.g., the original sequence is x ₁ , x ₂ ,..., x _n The difference is (x) ₂ -x ₁ ),(x ₃ -x ₂ ),...(x _n -x _n-1 ) In this way, the processing method is a common algorithm in the art, and is not described herein again. />

For the data at the time t-1 in the sequence with stationarity obtained after the difference, if the original sequence has stationarity, then->

Is x _t-1 。

In the training process of the integrated moving average autoregressive submodel, training data can be divided into a training set and a testing set, an AutoARIMA algorithm is adopted, a group of optimal p values and q values are determined on the training set, so that fitting of an ARIMA model can be achieved, the average absolute percentage error MAPE on the testing set is calculated, and the model with the minimum error is taken as the integrated moving average autoregressive submodel.

By way of example, may beDetermining a predicted feature parameter by the following formula

：

Wherein the content of the first and second substances,

is used to indicate the corresponding weight coefficient for the first prediction parameter, <' > is>

And is used for representing the weight coefficient corresponding to the second prediction parameter. />

For representing the error term. After the fitting of the time series prediction sub-model and the integrated moving average autoregressive sub-model is realized based on the above manner, the number group is calculated and obtained for the nearest N moments t based on the two fitted models>

And forming a data set>

Wherein is present>

Is the actual value corresponding to the sequence under the characteristic dimension.

In the training process, a training data feature sequence corresponding to the target object in each feature dimension may be determined based on the usage state data of the target object in the training period, where the training data feature sequence for model training may be generated by the above-described manner of generating the feature sequence data. For example, for a parameter prediction model corresponding to the feature of an active user, the parameter prediction model corresponding to the feature of the active user may be obtained by performing training fitting on the feature sequence of training data corresponding to the feature of the active user, and for a feature of an object change cycle, the parameter prediction model corresponding to the feature of the object change cycle may also be obtained by performing training fitting on the feature sequence of training data corresponding to the feature of the object change cycle. For the type replacement flow direction feature, training fitting may also be performed on the type replacement flow direction feature based on a training data feature sequence corresponding to the type replacement flow direction feature, so as to obtain a parameter prediction model corresponding to the type replacement flow direction feature. In the output of the parameter prediction model corresponding to the type change flow direction characteristic, if the sum of the prediction conversion rates for the same outflow type is 1, normalization processing may be performed after determining the conversion parameters for the same outflow type to each inflow type to obtain the prediction conversion rate for the outflow type to each inflow type, and a matrix formed by the prediction conversion rates for each outflow type to each inflow type is used as the prediction characteristic parameter corresponding to the type change flow direction characteristic. The parameter prediction model under each feature dimension can be obtained by training in the manner described above, and after the training is completed, the prediction feature parameters corresponding to the feature data sequences under each feature dimension can be determined according to the parameter prediction model corresponding to each feature dimension, so that the prediction of the prediction feature parameters under the active user features, the object replacement cycle features and the type replacement flow direction features is realized.

Therefore, the fusion sub-model can be trained on the basis of the data set, so that the weight coefficients respectively corresponding to the fusion of the first prediction parameter and the second prediction parameter in the fusion sub-model are obtained, the accurate fusion sub-model is obtained, and finally the parameter prediction model is obtained. In the parameter prediction model, the Prophet model predicts more accurately on a long sequence, and the ARIMA model predicts more accurately on a short sequence.

Embodiments of the present disclosure are described below with reference to the examples described in fig. 2 and 3. In this embodiment, the device identification codes of the target sub-objects may be identified, so that the types corresponding to the target sub-objects and the number of the target sub-objects in each type may be determined as the number of users of the target object in the type.

Further, a feature data sequence corresponding to each feature dimension corresponding to the target object may be respectively determined based on the usage state data of the target object in the current corresponding historical period, where the feature dimension may include an active user feature, an object replacement cycle feature, and a type replacement flow direction feature, and then the active user feature sequence, the object replacement cycle feature sequence, and the type replacement flow direction feature sequence may be respectively obtained. For example, the current corresponding historical period may be the last year, and the sequence of active user characteristics may include active user characteristics for the target object every month during the last year. Then, a prediction characteristic parameter corresponding to the active user characteristic can be determined based on the active user characteristic sequence and a parameter prediction model corresponding to the active user characteristic, a prediction characteristic parameter corresponding to the object change cycle characteristic can be determined based on the object change cycle characteristic sequence and a parameter prediction model corresponding to the object change cycle characteristic, and a prediction characteristic parameter corresponding to the type change flow direction characteristic can be determined based on the type change flow direction characteristic sequence and a parameter prediction model corresponding to the type change flow direction characteristic.

Taking type a in fig. 2 as an example of a target type, based on a product of the number of users of the current target object under type a and the predicted characteristic parameter corresponding to the predicted active user characteristic, it is determined that the current active user number of type a is 25494, which represents the current active user number of type a. Then, by dividing the active user number 25494 by the predicted feature parameter (i.e. 46.8) corresponding to the predicted object replacement cycle feature and multiplying by 12 months, it is determined that the predicted replacement number corresponding to the next year is 6537, i.e. the number of target objects that type a may replace in the next year. Further, according to the predicted feature parameters corresponding to the type replacement flow direction features predicted in fig. 3, the demand amount of the target object of type a generated when each type is replaced is determined. If the predicted replacement amount of type a in the next year is 6537, which is predicted from fig. 3 to indicate that there may be 38.0% of users to replace type a, the product of the predicted replacement amount 6537 and the predicted conversion rate 38.0% is determined as the amount of demand for the target objects of type a generated when the target objects of type a are replaced in the next year. In the same manner as described above, the demand amounts for the target objects of type a, which are generated when the target objects of type B, C, D, E, F, G are replaced in the next year, respectively, can be further determined, and the sum of the demand amounts for the target objects of type a, which are generated when the target objects of each type are replaced in the next year, is taken as the demand amount prediction data for the target objects of type a.

In a possible embodiment, before the step of determining the demand forecasting data of the target object according to the forecasting characteristic parameter and the number of users of the target object, the method may further include:

In an actual application scenario, when determining the demand of the target object, the situation that the training set of the model does not appear due to the current market environment factors may occur, for example, the capacity of type a is doubled, the capacity of type B is reduced, and the demand of the target object in different types is also affected. When demand prediction is performed through a machine learning model in the related art, the method is difficult to adapt to special situations in the scene. Based on this, in the embodiments of the present disclosure, adaptation to a particular scene may be performed by receiving an adjustment parameter.

For example, the target feature dimension may be a feature dimension influenced by the current market environment, which is determined based on expert experience, and the adjustment parameter may be determined for the current market environment, which indicates a final target for predicting the adjustment of the feature parameter. As an example, the adjustment parameter may be a parameter entered by a user through a parameter adjustment interface. As another example, an adjustment parameter corresponding to a preset scenario in at least one feature dimension may be preset, for example, when the capacity corresponding to a type decreases to a predetermined threshold, the predicted conversion rate of each outflow type to the type is set to 0.3. As an example, the predicted capacity of the target object under each type may be determined based on the prediction of the characteristics, such as raw materials, production equipment, and production capacity, required by the target object under each type, and if the predicted capacity of the type a drops to a predetermined threshold, the input of an adjustment parameter may be triggered, where a characteristic dimension corresponding to the adjustment parameter is a target characteristic dimension, and as described above, the type change of the adjustment parameter under the flow direction characteristic may be triggered, that is, the target average value of the predicted conversion rate of each outflow type to the type a is 0.3.

As an example, the adjusting the predicted feature parameter under the target feature dimension according to the adjustment parameter may include:

Wherein the target feature dimension may be a type change flow direction feature, such as a conversion of type a to type B needs to be adjusted to 0.2 due to the association between type a and type B. In the above example, if the predicted characteristic parameters are predicted, and the predicted conversion rate is 0.16 when the type a is changed to the type B in the predicted characteristic parameters, 1.25 (0.2/0.16) may be used as the adjustment ratio, and the predicted conversion rate 0.16 may be further adjusted to 0.2.

As another example, when the predicted characteristic parameters are predicted, the parameters corresponding to each month in the next quarter are predicted, and if the predicted conversion rates of changing the type a to the type B in the predicted characteristic parameters are 0.16, 0.14, and 0.18, respectively, and the average value of the predicted characteristic parameters is 0.16, the adjustment ratio is determined to be 1.25, and the predicted characteristic parameters obtained by adjusting based on the adjustment ratio are 0.2, 0.175, and 0.225, and further, the calculation of the demand prediction data of the subsequent target object can be performed based on the adjusted predicted characteristic parameters.

Therefore, according to the technical scheme, the prediction characteristic parameters under the characteristic dimension can be adjusted according to the actual condition corresponding to the target object in the process of determining the demand prediction data of the target object, so that the accuracy and the matching degree of the prediction characteristic parameters can be guaranteed by fusing a real application scene and expert knowledge, the adjustment flexibility in the process of determining the demand prediction data of the target object can be improved, the accuracy of the finally determined demand prediction data can be guaranteed, and accurate data support can be provided for inventory allocation or production guidance and the like based on the demand prediction data.

In one possible embodiment, the method may further comprise:

and acquiring the corresponding distribution proportion of the target object under a plurality of warehousing platforms.

The target object can correspond to different warehousing platforms in different sales scenes, and for example, the on-line demand and the off-line demand for the target object can be supplied through different warehousing platforms. Accordingly, the current distribution ratio may be predicted according to the online sales and offline sales in the historical period, wherein the prediction manner of the distribution ratio may be determined based on the prediction manner of a time series model commonly used in the art, such as an autoregressive model, a moving average model, and the like, which is not limited herein.

For another example, different sales platforms may correspond to different warehousing platforms, for example, the target object may be sold in a plurality of online sales platforms, and similarly, the allocation ratio of the warehousing platform corresponding to each sales platform may also be determined based on the sales amount corresponding to the different sales platforms in the historical period.

In this embodiment, the distribution can be performed according to the demand forecast data and the distribution proportion, so that the demand distribution quantity under each warehousing platform is matched with the sales volume corresponding to the warehousing platform, thereby avoiding the delivery delay caused by insufficient supply quantity to a certain extent and ensuring the user experience.

The present disclosure also provides a data processing apparatus, as shown in fig. 5, the data processing apparatus 10 including:

a first obtaining module 100 configured to obtain usage status data of a target object in a history period;

a first determining module 200 configured to determine a feature data sequence corresponding to the usage state data in each feature dimension;

a second determining module 300, configured to determine, for each of the feature dimensions, a predicted feature parameter corresponding to the feature data sequence in the feature dimension according to a parameter prediction model corresponding to the feature dimension;

a third determining module 400 configured to determine demand prediction data of the target object according to the prediction feature parameter and the number of users of the target object.

the third determining module comprises:

a second determining submodule configured to determine, according to the number of users of the target object in each of the association types and a predicted feature parameter corresponding to an active user feature in the association type, the number of active users of the target object in the association type, and determine, according to the number of active users of the target object in the association type and a predicted feature parameter corresponding to an object replacement cycle feature in the association type, a predicted replacement number of the target object in the association type;

Optionally, the second determining submodule is further configured to:

the second determining module includes:

Optionally, the apparatus further comprises:

Optionally, the adjusting module includes:

Optionally, the apparatus further comprises:

and the fourth determination module is configured to determine the distribution quantity of the demand quantity of the target object under each warehousing platform according to the demand quantity prediction data of the target object and the distribution proportion.

With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

The present disclosure also provides a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the steps of the data processing method provided by the present disclosure.

Fig. 6 is a block diagram illustrating an apparatus 800 for data processing in accordance with an example embodiment. For example, the apparatus 800 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, an exercise device, a personal digital assistant, and the like.

Referring to fig. 6, the apparatus 800 may include one or more of the following components: a first processing component 802, a first memory 804, a first power component 806, a multimedia component 808, an audio component 810, a first input/output interface 812, a sensor component 814, and a communication component 816.

The first processing component 802 generally controls overall operation of the device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The first processing component 802 may include one or more first processors 820 to execute instructions to perform all or a portion of the steps of the data processing method described above. Further, the first processing component 802 may include one or more modules that facilitate interaction between the first processing component 802 and other components. For example, the first processing component 802 may include a multimedia module to facilitate interaction between the multimedia component 808 and the first processing component 802.

The first memory 804 is configured to store various types of data to support operations at the apparatus 800. Examples of such data include instructions for any application or method operating on device 800, contact data, phonebook data, messages, pictures, videos, and so forth. The first memory 804 may be implemented by any type or combination of volatile or non-volatile memory devices, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.

A first power supply component 806 provides power to the various components of the device 800. The first power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the apparatus 800.

The multimedia component 808 includes a screen that provides an output interface between the device 800 and a user. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front facing camera and/or a rear facing camera. The front camera and/or the rear camera may receive external multimedia data when the device 800 is in an operation mode, such as a photographing mode or a video mode. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a Microphone (MIC) configured to receive external audio signals when the apparatus 800 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may further be stored in the first memory 804 or transmitted via the communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

The first input/output interface 812 provides an interface between the first processing component 802 and a peripheral interface module, which may be a keyboard, click wheel, button, etc. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.

The sensor assembly 814 includes one or more sensors for providing various aspects of state assessment for the device 800. For example, the sensor assembly 814 may detect the open/closed state of the device 800, the relative positioning of components, such as a display and keypad of the device 800, the sensor assembly 814 may also detect a change in position of the device 800 or a component of the device 800, the presence or absence of user contact with the device 800, the orientation or acceleration/deceleration of the device 800, and a change in temperature of the device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of a nearby object in the absence of any physical contact. The sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 816 is configured to facilitate communication between the apparatus 800 and other devices in a wired or wireless manner. The device 800 may access a wireless network based on a communication standard, such as WiFi,2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the apparatus 800 may be implemented by one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components for performing the above-described data processing methods.

In an exemplary embodiment, a non-transitory computer readable storage medium comprising instructions, such as the first memory 804 comprising instructions, executable by the first processor 820 of the apparatus 800 to perform the data processing method described above is also provided. For example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

The apparatus may be a part of a stand-alone electronic device, for example, in an embodiment, the apparatus may be an Integrated Circuit (IC) or a chip, where the IC may be one IC or a collection of multiple ICs; the chip may include, but is not limited to, the following categories: a GPU (Graphics Processing Unit), a CPU (Central Processing Unit), an FPGA (Field Programmable Gate Array), a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an SOC (System on Chip, SOC, system on Chip, or System on Chip), and the like. The integrated circuit or chip can be used to execute executable instructions (or codes) to realize the data processing method. Where the executable instructions may be stored in the integrated circuit or chip or may be retrieved from another device or apparatus, such as an integrated circuit or chip that includes a processor, memory, and an interface for communicating with other devices. The executable instructions may be stored in the memory, and when executed by the processor, implement the data processing method described above; alternatively, the integrated circuit or chip may receive executable instructions through the interface and transmit the executable instructions to the processor for execution, so as to implement the data processing method.

In another exemplary embodiment, a computer program product is also provided, which comprises a computer program executable by a programmable apparatus, the computer program having code portions for performing the above-mentioned data processing method when executed by the programmable apparatus.

Fig. 7 is a block diagram illustrating an apparatus 1900 for data processing according to an example embodiment. For example, the apparatus 1900 may be provided as a server. Referring to FIG. 7, apparatus 1900 includes a second processing component 1922 further including one or more processors and memory resources represented by a second memory 1932 for storing instructions, e.g., applications, executable by second processing component 1922. The application programs stored in the second memory 1932 may include one or more modules that each correspond to a set of instructions. Further, the second processing component 1922 is configured to execute instructions to perform the data processing method described above.

The device 1900 may also include a second power component 1926 configured to perform power management of the device 1900, a wired or wireless network interface 1950 configured to connect the device 1900 to a network, and a second input/output interface 1958. The device 1900 may operate based on an operating system, such as Windows Server, stored in memory 1932 ^TM ，Mac OS X ^TM ，Unix ^TM ，Linux ^TM ，FreeBSD ^TM Or the like.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A data processing method, comprising:

acquiring use state data of a target object in a historical period;

determining a characteristic data sequence corresponding to the use state data in each characteristic dimension;

2. The method of claim 1, wherein the target object corresponds to a plurality of types, and the feature dimensions include an active user feature, an object replacement cycle feature, and a type replacement flow direction feature for each of the types;

the determining demand forecast data of the target object according to the forecast characteristic parameter and the user number of the target object comprises the following steps:

3. The method according to claim 2, wherein the determining the predicted replacement number of the target object in the association type according to the active user number of the target object in the association type and a predicted feature parameter corresponding to a feature of a replacement cycle of the target object in the association type comprises:

4. The method of claim 1, wherein the parametric prediction model comprises a time series predictor model, an integrated moving average autoregressive submodel, and a fusion submodel;

5. The method of claim 1, wherein prior to the step of determining demand prediction data for the target object based on the predicted characteristic parameter and the number of users of the target object, the method further comprises:

6. The method according to claim 5, wherein the adjusting the predicted feature parameters in the target feature dimension according to the adjustment parameters comprises:

7. The method of claim 1, further comprising:

8. The method of claim 1, wherein the feature dimensions comprise active user features; a plurality of target sub-objects are correspondingly arranged under the target object, and each target sub-object is electronic equipment provided with an application program;

determining the operating parameters of each target sub-object under the target object at each sampling moment in the historical time period according to the use state data of the target object, wherein the operating parameters comprise the use frequency of a target application program and/or the screen use duration ratio of the target sub-object;

9. A data processing apparatus, characterized by comprising:

10. A data processing apparatus, comprising:

a processor;

a memory for storing processor-executable instructions;

wherein the processor is configured to:

acquiring use state data of a target object in a historical period;

11. A computer-readable storage medium, on which computer program instructions are stored, which program instructions, when executed by a processor, carry out the steps of the method according to any one of claims 1 to 8.