CN112116381A - Moon life prediction method based on LSTM neural network, storage medium and computer equipment - Google Patents

Moon life prediction method based on LSTM neural network, storage medium and computer equipment Download PDF

Info

Publication number
CN112116381A
CN112116381A CN202010894199.2A CN202010894199A CN112116381A CN 112116381 A CN112116381 A CN 112116381A CN 202010894199 A CN202010894199 A CN 202010894199A CN 112116381 A CN112116381 A CN 112116381A
Authority
CN
China
Prior art keywords
day
sequence
neural network
month
lstm neural
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010894199.2A
Other languages
Chinese (zh)
Other versions
CN112116381B (en
Inventor
李慧斌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Keynote Network Inc
Original Assignee
Beijing Keynote Network Inc
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 Beijing Keynote Network Inc filed Critical Beijing Keynote Network Inc
Priority to CN202010894199.2A priority Critical patent/CN112116381B/en
Publication of CN112116381A publication Critical patent/CN112116381A/en
Application granted granted Critical
Publication of CN112116381B publication Critical patent/CN112116381B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06Q30/0202—Market predictions or forecasting for commercial activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The present application relates to a method, storage medium and computer device for predicting monthly activity based on an LSTM neural network, wherein the method comprises: receiving the accumulated active equipment number of each day from the ith day to the tth day to obtain a first sequence; performing differential operation on the first sequence to obtain a second sequence, wherein the second sequence is the number of newly added active devices from the i +1 th day to the t th day; determining the number of newly added active devices each day from the t +1 th day to the end of the month by using an LSTM neural network according to the second sequence; and determining the monthly active equipment number according to the newly increased daily active equipment number from the t +1 th day to the end of the month and the accumulated active equipment number of each day from the ith day to the tth day. By the method and the device, the number of active devices in the current month is predicted, and the method and the device do not depend on the number of active devices in the historical month and are not influenced by artificial priori-based meanings.

Description

Moon life prediction method based on LSTM neural network, storage medium and computer equipment
Technical Field
The application relates to the technical field of data analysis, in particular to a monthly life prediction method based on a Long Short-Term Memory (LSTM) neural network, a storage medium and computer equipment.
Background
In the related art, the number of monthly active devices (abbreviated as monthly activity) is predicted based on historical monthly activity. For example, a model is used for modeling the historical monthly activity of an application, the model is ARIMA, polynomial fitting, exponential averaging, and the like, and the monthly activity of the next month is predicted according to the historical monthly activity. The method cannot realize rapid prediction depending on self historical data.
In the related art, the prediction of the monthly activity is also based on the daily activity and the ratio of the daily activity to the monthly activity in the same type of application. For example, the daily activity d of a certain e-commerce App is counted, and the daily activities d 'of apps of the same kind are counted, and meanwhile, the monthly activity m' of apps of the same kind is known, then the monthly activity m ═ d (m '/d') of the e-commerce App is estimated, and in order to further increase the precision, d can be an average value of the last n days, and d 'and m' can be statistical data of a plurality of apps of the same kind. Although the above method solves the problem of fast estimation, sometimes we sort by manual methods based on a priori policy is not always reasonable, for example, although also for e-commerce, some App users have high and low repeated login times.
Disclosure of Invention
To solve the above technical problem or at least partially solve the above technical problem, the present application provides a monthly life prediction method based on an LSTM neural network, a storage medium, and a computer device.
In a first aspect, the present application provides a method for predicting monthly activity based on an LSTM neural network, including: receiving the accumulated active equipment number of each day from the ith day to the tth day to obtain a first sequence; performing differential operation on the first sequence to obtain a second sequence, wherein the second sequence is the number of newly added active devices from the i +1 th day to the t th day; determining the number of newly added active devices each day from the t +1 th day to the end of the month by using an LSTM neural network according to the second sequence; and determining the monthly active equipment number according to the newly increased daily active equipment number from the t +1 th day to the end of the month and the accumulated active equipment number of each day from the ith day to the tth day.
In some embodiments, before determining the number of newly added active devices per day from day t +1 to the end of the month using the LSTM neural network according to the second sequence, the method further comprises: normalizing the second sequence by using the maximum value in the second sequence to obtain a third sequence; and determining the number of newly added active devices per day from the t +1 th day to the end of the month according to a second sequence using the LSTM neural network, comprising: determining a fourth sequence according to the third sequence by using an LSTM neural network, wherein the fourth sequence corresponds to the number of newly increased active devices each day from the t +1 th day to the end of the month; and performing inverse normalization processing on the fourth sequence by using the maximum value in the second sequence to obtain the number of newly increased active devices each day from the t +1 th day to the end of the month.
In some embodiments, determining the monthly active device count based on the daily additional active device count from day t +1 to the end of the month and the cumulative active device count for each day from day i to day t includes: and determining the sum of the accumulated active equipment number of the t th day and the newly increased active equipment number of the t +1 th day to the end of the month, and taking the determination result as the monthly active equipment number.
In some embodiments, determining said fourth sequence from said third sequence using an LSTM neural network comprises: and using an LSTM neural network, obtaining an output of a third sequence by taking the third sequence as an input, adding the output into the third sequence to be used as a new third sequence, and circulating until an output corresponding to the end of the month is obtained to obtain a fourth sequence, wherein the fourth sequence corresponds to the number of newly increased active devices per day from the t +1 th day to the end of the month.
In some embodiments, determining the number of newly added active devices per day from day t +1 to the end of the month using the LSTM neural network according to the second sequence comprises: and determining the number of the newly increased active devices per day on the t +1 th day by using the LSTM neural network and taking the second sequence as input, adding the number of the newly increased active devices per day on the t +1 th day into the second sequence to obtain a new second sequence, and circulating until the number of the newly increased active devices per day at the end of the month is obtained.
In some embodiments, the LSTM neural network is trained with the number of newly added active devices per day from day 1 to day n-1 of the historical month as input and with the number of newly added active devices per day from day 2 to day n of the historical month as expected.
In some embodiments, the LSTM neural network is trained with the normalized sequence of the number of newly added active devices per day from day 1 to day n-1 of the historical month as input and with the normalized sequence of the number of newly added active devices per day from day 2 to day n of the historical month as expected.
In certain embodiments, the LSTM neural network is trained using AUC (Area Under cutter, defined as the Area Under the receiver operating characteristic Curve (ROC Curve) with respect to coordinate axes) or Mean Square Error (MSE) as a model training metric.
In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor; the computer program, when executed by a processor, performs the steps of a method for predicting monthly activity based on an LSTM neural network.
In a third aspect, the present application provides a computer readable storage medium, on which an LSTM neural network-based month activity prediction program is stored, the LSTM neural network-based month activity prediction program implementing the steps of the LSTM neural network-based month activity prediction method when executed by a processor.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages: the method provided by the embodiment of the application realizes the prediction of the number of active devices in the current month, does not depend on the number of active devices in the historical month, and is not influenced by artificial prior.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a flow chart of one embodiment of a method for predicting monthly activity based on an LSTM neural network provided herein;
FIG. 2 is a flow chart of another embodiment of a method for predicting monthly activity based on an LSTM neural network provided in the present application;
FIG. 3 is a schematic diagram of an embodiment of a system according to the present disclosure;
fig. 4 is a block diagram illustrating a structure of an implementation manner of a prediction apparatus 340 according to an embodiment of the present disclosure; and
fig. 5 is a hardware structure diagram of an implementation manner of a computer device according to an embodiment of the present application.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In the following description, suffixes such as "module", "component", or "unit" used to denote elements are used only for facilitating the explanation of the present invention, and have no specific meaning in itself. Thus, "module", "component" or "unit" may be used mixedly.
Fig. 1 is a flowchart of an embodiment of a method for predicting monthly activity based on an LSTM neural network provided in the present application, as shown in fig. 1, the method includes steps S102 to S108.
Step S102, receiving the accumulated active equipment number of each day from the ith day to the tth day to obtain a first sequence.
Herein, the cumulative active device count for each day is the active device count counted for each day, and the cumulative active device count at the end of the month is the active device count for the month of the month. The first sequence is shown in table 1.
TABLE 1 first sequence (cumulative number of active devices per day) schematic Table
Date 1 2 … n
Accumulating active device count x1 x2 … xn
Wherein n is the number of days of the month. From day i to day t is { xi，…，xtAnd the set comprises t-i +1 elements.
And step S104, performing difference operation on the first sequence to obtain a second sequence, wherein the second sequence is the number of newly added active devices from the i +1 th day to the t th day.
Herein, a second sequence is determined based on the first sequence, the second sequence being the number of newly added active devices per day from day i +1 to day t. The second sequence is shown in table 2.
TABLE 2 second sequence (number of newly added active devices per day) schematic Table
Date 2 3 … n
Difference (diff) x2-x1 x3-x2 … xn-xn-1
Wherein n is the number of days of the month. The number of newly added active devices per day from day i +1 to day t is { xi+1-xi，。。。，xt–xt-1And the set comprises t-i elements.
And step S106, determining the number of newly added active devices from the t +1 th day to the end of the month according to the second sequence by using the LSTM neural network.
And step S108, determining the monthly active equipment number according to the newly increased daily active equipment number from the t +1 th day to the end of the month and the accumulated active equipment number of each day from the ith day to the tth day.
In this embodiment of the application, the cumulative number of active devices in each day from the i-th day to the t-th day is known, and the number of newly-added active devices in each day from the t + 1-th day to the end of the month is predicted, and step S108 may determine the number of active devices in each month according to the known cumulative number of active devices and the predicted number of newly-added active devices in each day thereafter.
In step S106, the number of newly added active devices per day from the t +1 th day to the end of the month is predicted by using LSTM, and the cumulative number of active devices per day from the i th day to the t th day is received in step S102, so that the number of active devices per month can be determined. In some embodiments, in step S108, the sum of the cumulative number of active devices on the t-th day and the number of newly-added active devices on each day from the t +1 th day to the end of the month is determined, and the determination result is used as the number of active devices on the month, but the present application is not limited thereto.
In some embodiments, the LSTM neural network is used to perform multi-step prediction, and in step S106, the LSTM neural network is used to determine the number of newly-added daily active devices for the t +1 th day using the second sequence as input, add the number of newly-added daily active devices for the t +1 th day into the second sequence to obtain a new second sequence, and circulate until the number of newly-added daily active devices for the end of the month is obtained.
In practical applications, the cumulative number of active devices for several days is easily obtained, for example, the cumulative number of active devices for several days can be collected after the APP is operated online for several days. Thus, the monthly activity of the APP can be quickly predicted with the APP only operating for a few days.
Fig. 2 is a flowchart of another embodiment of the method for predicting monthly activity based on the LSTM neural network according to the present application, in which normalization processing is performed on data to reduce errors and improve prediction accuracy, as shown in fig. 2, the method for predicting monthly activity includes steps S202 to S212.
Step S202, receiving the accumulated active equipment number of each day from the ith day to the tth day to obtain a first sequence.
Herein, the cumulative active device count for each day is the active device count counted for each day, and the cumulative active device count at the end of the month is the active device count for the month of the month. The first sequence is shown in table 1.
And step S204, carrying out difference operation on the first sequence to obtain a second sequence, wherein the second sequence is the number of newly added active devices from the i +1 th day to the t th day.
In step S204, the second sequence is shown in table 2.
And step S206, normalizing the second sequence by using the maximum value in the second sequence to obtain a third sequence. The third sequence is shown in table 3.
TABLE 3 third sequence schematic Table
Date 2 3 … n
Normalization (x2-x1)/maxdiff (x3-x2)/maxdiff … (xn-xn-1)/maxdiff
Therein, maxdiffIs the maximum value in the second sequence. The third sequence from day i +1 to day t is { (x)i+1-xi)/maxdiff，…，(xt–xt-1)/maxdiffAnd the set comprises t-i elements.
And S208, determining a fourth sequence according to the third sequence by using the LSTM neural network, wherein the fourth sequence corresponds to the number of newly added active devices each day from the t +1 th day to the end of the month.
Step S210, using the maximum value in the second sequence (see max in Table 3)diff) And performing inverse normalization processing on the fourth sequence to obtain the number of newly increased active equipment each day from the t +1 th day to the end of the month.
Herein, elements in the fourth sequence are denoted scalesiThe fourth sequence determined in step S208 is { scale }t+1，…，scalen}。
And step S212, determining the monthly active equipment number according to the newly increased daily active equipment number from the t +1 th day to the end of the month and the accumulated active equipment number of each day from the ith day to the tth day.
In some embodiments, the LSTM neural network is used to perform the multi-step prediction, and in step S208, the LSTM neural network is used to obtain an output of a third sequence using the third sequence as an input, add the output to the third sequence as a new third sequence, and circulate until an output corresponding to the end of the month is obtained to obtain the fourth sequence, where the fourth sequence corresponds to the number of newly added active devices per day from the t +1 th day to the end of the month.
In some embodiments, in the step S212, the sum of the cumulative number of active devices on the t th day and the number of newly added active devices each day from the t +1 th day to the end of the month is determined, and the determination result is used as the number of active devices each day, but the present application is not limited thereto, and a person skilled in the art may perform a certain process (for example, equal weighted average) on the cumulative number of active devices on the t th day and the number of newly added active devices each day from the t +1 th day to the end of the month, and any method based on the known cumulative number of active devices and the predicted number of newly added active devices each day thereafter is feasible.
Fig. 3 is a schematic structural diagram of an implementation manner of a system provided in an embodiment of the present application, and as shown in fig. 3, the system includes: user equipment 310, server 320, database 330, prediction means 340 and application means 350. Therein, the service 320 is communicatively connected with the user device 310 for collecting the active device number, e.g. by counting the active device number by a device Identification (ID) of the user device 310 using an Application (APP). Database 330 is set to store the number of active devices collected by server 320. The predicting means 340 is configured to predict the monthly active device number based on the collected active device numbers. The application means 350 is configured to apply the monthly active equipment number predicted by the prediction means 340, such as calculating a sampling rate to reduce storage and loan costs, dynamically expanding capacity, generating usage reports, making usage restrictions, and the like, but is not limited thereto.
It should be understood that fig. 3 is only an exemplary illustration, and active device data may be collected in any manner before or after the filing date of the present application, which is not limited by the embodiments of the present application.
Fig. 4 is a block diagram of a configuration of an implementation manner of a prediction apparatus 340 according to an embodiment of the present application, and as shown in fig. 4, the prediction apparatus 340 includes: a receiving module 341 configured to receive the cumulative number of active devices from the ith day to the tth day to obtain a first sequence; a first determining module 342, connected to the receiving module 341, configured to perform a difference operation on the first sequence to obtain a second sequence, where the second sequence is the number of newly added active devices each day from the i +1 th day to the t-th day; a second determining module 343, connected to the first determining module 342, configured to determine, according to the second sequence, the number of newly added active devices per day from the t +1 th day to the end of the month using the LSTM neural network; the third determining module 344 determines the monthly active device count according to the number of newly added active devices each day from the t +1 th day to the end of the month and the cumulative active device count for each day from the i-th day to the t-th day.
In some embodiments, referring to fig. 4, the predicting apparatus 340 further includes a fourth determining module 345 configured to normalize the second sequence by using the maximum value in the second sequence to obtain a third sequence. The third determining module 344 is configured to determine a fourth sequence according to the third sequence by using the LSTM neural network, where the fourth sequence corresponds to the number of newly added active devices each day from the t +1 th day to the end of the month. The prediction apparatus 340 further comprises a fifth determining module 346 configured to perform inverse normalization on the fourth sequence by using the maximum value in the second sequence to obtain the number of newly added active devices per day from the t +1 th day to the end of the month.
In this embodiment of the application, the cumulative number of active devices in each day from the i-th day to the t-th day is known, and the number of newly-added active devices in each day from the t + 1-th day to the end of the month is predicted, and the third determining module 344 is configured to determine the number of active devices in each month according to the known cumulative number of active devices and the predicted number of newly-added active devices in each day thereafter.
LSTM neural network training
The training process of the LSTM neural network can be found in the related art. The input data of the LSTM neural network is exemplified in the embodiments of the present application.
In the embodiment of the application, a prediction method for predicting the monthly activity data by the monthly activity data in the related art is changed. In some embodiments, the LSTM neural network is trained with the number of newly added active devices per day from day 1 to day n-1 of the historical month as input and with the number of newly added active devices per day from day 2 to day n of the historical month as expected. Therefore, the number of the newly added active devices per day predicted by the LSTM application network is close to the actual number of the newly added active devices per day.
In some embodiments, the LSTM neural network is trained with the normalized sequence of the number of newly added active devices per day from day 1 to day n-1 of the historical month as input and with the normalized sequence of the number of newly added active devices per day from day 2 to day n of the historical month as expected, such that the LSTM predicts unknown data from the known number of newly added active devices per day for each day.
In some embodiments, the LSTM neural network is trained using AUC (Area Under cutter, defined as the Area enclosed by the axes Under the ROC Curve) or Mean Square Error (MSE) as a model training metric. The training process using AUC or MSE as the model training index is referred to in the known art before the filing date of the present application, which is not described in detail in the embodiments of the present application, and the implementation of the present application is not limited thereto.
In some embodiments, the LSTM neural network is trained to predict the monthly active device count for a plurality of APPs (but not limited to) months (but not limited to) using the cumulative active device count for the plurality of APPs (but not limited to) days as a training data set for the LSTM neural network. As an exemplary illustration, the cumulative number of active devices in the training data set is preprocessed by taking a month share as a unit, and the cumulative number of active devices in each day of the month is differentially processed to obtain the number of newly added active devices in each day of the month. And normalizing the number of the newly increased active devices every day in the current month by using the maximum number of the newly increased active devices every day in the current month to obtain a normalization sequence in the current month. This results in a normalized sequence of a number of APP months. The LSTM neural network is trained with a normalized sequence of a number of APP months as the data input.
In some embodiments, the LSTM neural network has a neuron number of 20 and a maximum step size of 32.
It should be understood that the training of the LSTM neural network is only exemplified herein, and in the embodiment of the present application, a sequence of the number of newly added active devices per day obtained by accumulating the number of active devices per day and the number of newly added active devices per month is feasible as the data input of the LSTM neural network.
The embodiment also provides a computer device, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack server, a blade server, a tower server or a rack server (including an independent server or a server cluster composed of a plurality of servers) capable of executing programs, and the like. The computer device 20 of the present embodiment includes at least, but is not limited to: a memory 21, a processor 22, which may be communicatively coupled to each other via a system bus, as shown in FIG. 5. It is noted that fig. 5 only shows a computer device 20 with components 21-22, but it is to be understood that not all shown components are required to be implemented, and that more or fewer components may be implemented instead.
In the present embodiment, the memory 21 (i.e., a readable storage medium) includes a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, and the like. In some embodiments, the storage 21 may be an internal storage unit of the computer device 20, such as a hard disk or a memory of the computer device 20. In other embodiments, the memory 21 may also be an external storage device of the computer device 20, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), or the like, provided on the computer device 20. Of course, the memory 21 may also include both internal and external storage devices of the computer device 20. In this embodiment, the memory 21 is generally used for storing an operating system installed in the computer device 20 and various application software, such as program codes of the LSTM neural network-based monthly life prediction method. Further, the memory 21 may also be used to temporarily store various types of data that have been output or are to be output.
Processor 22 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data Processing chip in some embodiments. The processor 22 is typically used to control the overall operation of the computer device 20. In this embodiment, the processor 22 is configured to execute the program code stored in the memory 21 or process data, such as the program code of the LSTM neural network-based monthly life prediction method, to implement the LSTM neural network-based monthly life prediction method.
The present embodiment also provides a computer-readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, a server, an App application mall, etc., on which a computer program is stored, which when executed by a processor implements corresponding functions. The computer-readable storage medium of the present embodiment is for storing an LSTM neural network-based monthly activity prediction program, and the steps of the LSTM neural network-based monthly activity prediction method are implemented when being executed by a processor.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (10)

1. A month activity prediction method based on a long-term and short-term memory LSTM neural network is characterized by comprising the following steps:
receiving the accumulated active equipment number of each day from the ith day to the tth day to obtain a first sequence;
performing differential operation on the first sequence to obtain a second sequence, wherein the second sequence is the number of newly added active devices from the i +1 th day to the t th day;
determining the number of newly added active devices each day from the t +1 th day to the end of the month by using an LSTM neural network according to the second sequence;
and determining the monthly active equipment number according to the newly increased daily active equipment number from the t +1 th day to the end of the month and the accumulated active equipment number of each day from the ith day to the tth day.
2. The LSTM neural network-based monthly activity prediction method of claim 1,
before determining the number of newly added active devices per day from the t +1 th day to the end of the month according to the second sequence by using the LSTM neural network, the method further comprises the following steps: normalizing the second sequence by using the maximum value in the second sequence to obtain a third sequence; and
determining, using the LSTM neural network, a daily number of newly added active devices from day t +1 to the end of the month according to the second sequence, comprising:
determining a fourth sequence according to the third sequence by using an LSTM neural network, wherein the fourth sequence corresponds to the number of newly added active devices each day from the t +1 th day to the end of the month; and
and performing inverse normalization processing on the fourth sequence by using the maximum value in the second sequence to obtain the number of newly increased active devices each day from the t +1 th day to the end of the month.
3. The LSTM neural network-based monthly activity prediction method of claim 1, wherein determining monthly active device counts based on the daily additional active device counts from the t +1 th day to the end of the month and the cumulative active device counts from the i th day to the t th day for each day comprises:
and determining the sum of the accumulated active equipment number of the t th day and the newly increased active equipment number of the t +1 th day to the end of the month, and taking the determination result as the monthly active equipment number.
4. The LSTM neural network-based monthly activity prediction method of claim 2, wherein determining a fourth sequence from the third sequence using an LSTM neural network comprises:
and using an LSTM neural network, taking the third sequence as input to obtain output of the third sequence, adding the output into the third sequence to be used as a new third sequence, and circulating until obtaining output corresponding to the end of the month to obtain a fourth sequence, wherein the fourth sequence corresponds to the number of newly increased active devices per day from the t +1 th day to the end of the month.
5. The LSTM neural network-based monthly activity prediction method of claim 1, wherein determining the number of newly added active devices per day from the t +1 th day to the end of the month using the LSTM neural network from the second sequence comprises:
and determining the number of newly increased active devices per day on the t +1 th day by using the LSTM neural network and taking the second sequence as input, adding the number of newly increased active devices per day on the t +1 th day into the second sequence to obtain a new second sequence, and circulating until the number of newly increased active devices per day at the end of the month is obtained.
6. The LSTM neural network-based monthly activity prediction method of claim 1 or 2, wherein the LSTM neural network is trained with the number of newly added active devices per day from day 1 to day n-1 of the historical month as input and with the number of newly added active devices per day from day 2 to day n of the historical month as expectation.
7. The LSTM neural network-based monthly activity prediction method of claim 2 or 4, wherein the LSTM neural network is trained with the normalized sequence of the number of newly added daily active devices from day 1 to day n-1 of the historical month as input and with the normalized sequence of the number of newly added daily active devices from day 2 to day n of the historical month as expected.
8. The LSTM neural network-based monthly life prediction method of claim 1, wherein the LSTM neural network is trained using an area AUC or mean square error MSE enclosed by coordinate axes under a working characteristic curve of a subject as a model training index.
9. A computer device, characterized in that the computer device comprises:
a memory, a processor, and a computer program stored on the memory and executable on the processor;
the computer program when executed by the processor implements the steps of the long-short term memory (LSTM) neural network based monthly activity prediction method of any one of claims 1 to 8.
10. A computer-readable storage medium, wherein the computer-readable storage medium stores thereon a long-short term memory LSTM neural network-based monthly activity prediction program, which when executed by a processor implements the steps of the LSTM neural network-based monthly activity prediction method of any one of claims 1-8.
CN202010894199.2A 2020-08-31 2020-08-31 Moon life prediction method based on LSTM neural network, storage medium and computer equipment Active CN112116381B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010894199.2A CN112116381B (en) 2020-08-31 2020-08-31 Moon life prediction method based on LSTM neural network, storage medium and computer equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010894199.2A CN112116381B (en) 2020-08-31 2020-08-31 Moon life prediction method based on LSTM neural network, storage medium and computer equipment

Publications (2)

Publication Number Publication Date
CN112116381A true CN112116381A (en) 2020-12-22
CN112116381B CN112116381B (en) 2021-05-07

Family

ID=73804546

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010894199.2A Active CN112116381B (en) 2020-08-31 2020-08-31 Moon life prediction method based on LSTM neural network, storage medium and computer equipment

Country Status (1)

Country Link
CN (1) CN112116381B (en)

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106600344A (en) * 2016-12-30 2017-04-26 广州虎牙信息科技有限公司 Method and apparatus for obtaining active user data of target product
CN108596652A (en) * 2018-03-28 2018-09-28 麒麟合盛网络技术股份有限公司 Active users prediction technique and device
CN108881333A (en) * 2017-05-09 2018-11-23 腾讯科技（深圳）有限公司 A kind of method and apparatus for predicting to enliven number of objects day
CN109063927A (en) * 2018-08-28 2018-12-21 成都信息工程大学 A kind of microblogging transfer amount prediction technique based on TS-LSTM and DNN
CN109194498A (en) * 2018-07-27 2019-01-11 南京理工大学 A kind of network flow prediction method based on LSTM
CN109360031A (en) * 2018-11-16 2019-02-19 北京奇虎科技有限公司 A kind of prediction technique and device of active users
CN109583946A (en) * 2018-11-16 2019-04-05 北京奇虎科技有限公司 A kind of forecasting system and method for active users
CN109784997A (en) * 2019-01-11 2019-05-21 重庆邮电大学 A kind of short-sighted frequency any active ues prediction technique based on big data
CN109858694A (en) * 2019-01-28 2019-06-07 上海连尚网络科技有限公司 A kind of method and apparatus of day active users prediction
CN110782096A (en) * 2019-10-29 2020-02-11 山东科技大学 Forex time series prediction method
CN110796484A (en) * 2019-10-11 2020-02-14 上海上湖信息技术有限公司 Method and device for constructing customer activity degree prediction model and application method thereof
EP3640902A2 (en) * 2018-10-17 2020-04-22 Tata Consultancy Services Limited System and method for authenticating humans based on behavioral pattern
CN111401433A (en) * 2020-03-12 2020-07-10 腾讯科技（深圳）有限公司 User information acquisition method and device, electronic equipment and storage medium

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106600344A (en) * 2016-12-30 2017-04-26 广州虎牙信息科技有限公司 Method and apparatus for obtaining active user data of target product
CN108881333A (en) * 2017-05-09 2018-11-23 腾讯科技（深圳）有限公司 A kind of method and apparatus for predicting to enliven number of objects day
CN108596652A (en) * 2018-03-28 2018-09-28 麒麟合盛网络技术股份有限公司 Active users prediction technique and device
CN109194498A (en) * 2018-07-27 2019-01-11 南京理工大学 A kind of network flow prediction method based on LSTM
CN109063927A (en) * 2018-08-28 2018-12-21 成都信息工程大学 A kind of microblogging transfer amount prediction technique based on TS-LSTM and DNN
EP3640902A2 (en) * 2018-10-17 2020-04-22 Tata Consultancy Services Limited System and method for authenticating humans based on behavioral pattern
CN109583946A (en) * 2018-11-16 2019-04-05 北京奇虎科技有限公司 A kind of forecasting system and method for active users
CN109360031A (en) * 2018-11-16 2019-02-19 北京奇虎科技有限公司 A kind of prediction technique and device of active users
CN109784997A (en) * 2019-01-11 2019-05-21 重庆邮电大学 A kind of short-sighted frequency any active ues prediction technique based on big data
CN109858694A (en) * 2019-01-28 2019-06-07 上海连尚网络科技有限公司 A kind of method and apparatus of day active users prediction
CN110796484A (en) * 2019-10-11 2020-02-14 上海上湖信息技术有限公司 Method and device for constructing customer activity degree prediction model and application method thereof
CN110782096A (en) * 2019-10-29 2020-02-11 山东科技大学 Forex time series prediction method
CN111401433A (en) * 2020-03-12 2020-07-10 腾讯科技（深圳）有限公司 User information acquisition method and device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN112116381B (en) 2021-05-07

Similar Documents

Publication Publication Date Title
CN110263152B (en) Text classification method, system and computer equipment based on neural network
CN109388674B (en) Data processing method, device, equipment and readable storage medium
EP3543941A1 (en) Time-factor fusion collaborative filtering method, device, server and storage medium
CN108229986B (en) Feature construction method in information click prediction, information delivery method and device
CN113255370B (en) Industry type recommendation method, device, equipment and medium based on semantic similarity
CN110599354B (en) Online checking method, online checking system, computer device and computer readable storage medium
CN112328909B (en) Information recommendation method and device, computer equipment and medium
CN108182633B (en) Loan data processing method, loan data processing device, loan data processing program, and computer device and storage medium
US20210089887A1 (en) Variance-Based Learning Rate Control For Training Machine-Learning Models
CN110807547A (en) Method and system for predicting family population structure
CN110880014A (en) Data processing method and device, computer equipment and storage medium
CN111797320A (en) Data processing method, device, equipment and storage medium
CN112819157A (en) Neural network training method and device and intelligent driving control method and device
CN113312578A (en) Data index fluctuation attribution method, device, equipment and medium
CN111626898B (en) Method, device, medium and electronic equipment for realizing attribution of events
CN112116381B (en) Moon life prediction method based on LSTM neural network, storage medium and computer equipment
CN112887371A (en) Edge calculation method and device, computer equipment and storage medium
CN111783883A (en) Abnormal data detection method and device
CN114880363A (en) Data center flow prediction system, training method and prediction method
CN109885710B (en) User image depicting method based on differential evolution algorithm and server
CN113780666A (en) Missing value prediction method and device and readable storage medium
CN112784165A (en) Training method of incidence relation estimation model and method for estimating file popularity
CN112667754A (en) Big data processing method and device, computer equipment and storage medium
CN112669091A (en) Data processing method, device and storage medium
CN109981327B (en) Method and system for predicting traffic

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant