CN112231548A - User login behavior analysis method, device and system and storage medium - Google Patents
User login behavior analysis method, device and system and storage medium Download PDFInfo
- Publication number
- CN112231548A CN112231548A CN201910636193.2A CN201910636193A CN112231548A CN 112231548 A CN112231548 A CN 112231548A CN 201910636193 A CN201910636193 A CN 201910636193A CN 112231548 A CN112231548 A CN 112231548A
- Authority
- CN
- China
- Prior art keywords
- user
- analyzing
- behavior data
- login
- purchase
- 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.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims description 42
- 230000006399 behavior Effects 0.000 claims abstract description 129
- 238000000034 method Methods 0.000 claims abstract description 66
- 238000006243 chemical reaction Methods 0.000 claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims description 44
- 238000012549 training Methods 0.000 claims description 29
- 230000007704 transition Effects 0.000 claims description 28
- 230000015654 memory Effects 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 6
- 230000000694 effects Effects 0.000 abstract description 11
- 238000004422 calculation algorithm Methods 0.000 description 18
- 238000007637 random forest analysis Methods 0.000 description 15
- 230000008569 process Effects 0.000 description 13
- 238000013461 design Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 11
- 230000009471 action Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 3
- 238000005192 partition Methods 0.000 description 3
- 238000007635 classification algorithm Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method, a device, a system and a storage medium for analyzing user login behaviors, wherein the method comprises the following steps: acquiring online behavior data of a user within a preset time period; analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data; constructing a hidden Markov model according to the purchase willingness value of the user in the future stage; and analyzing the login condition of the user in the future stage through the hidden Markov model. The method and the device can analyze the login condition of the future stage according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the future stage of the user, the conversion effect of the recommendation strategy is improved, and the user experience is good.
Description
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a method, an apparatus, a system, and a storage medium for analyzing a user login behavior.
Background
More and more user data on the internet are recorded, and the action data of searching, clicking, purchasing and the like of the user on a website reflect the interest and the demand of the user on products. By analyzing and modeling a series of behavior data of the user, the degree of interest of the user in products and websites can be analyzed, so that data support is provided for user recommendation, website popularization and the like.
At present, a deep learning model is generally combined to analyze massive online data of a user, and then commodities which the user may order and purchase are analyzed.
However, this method can only analyze the purchasing behavior of the logged-in user, and it is difficult to analyze the logging-in behavior of the user at a future stage, so that the user who may log in cannot be recommended in advance, and the recommendation policy effect is poor.
Disclosure of Invention
The invention provides an analysis method, device and system for user login behaviors and a storage medium, which can analyze the login situation of a future stage according to the online behavior data of a user, so that accurate recommendation can be performed on the user according to the login situation of the user at the future stage, the conversion effect of a recommendation strategy is improved, and the user experience is good.
In a first aspect, an embodiment of the present invention provides a method for analyzing a user login behavior, including:
acquiring online behavior data of a user within a preset time period;
analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data;
constructing a hidden Markov model according to the purchase willingness value of the user in the future stage;
and analyzing the login condition of the user in the future stage through the hidden Markov model.
In one possible design, the inline behavior data includes: the method comprises the following steps of commodity browsing record, ordering record, change record of commodities in a shopping cart and attention commodity record.
In one possible design, further comprising:
constructing a training sample through historical online behavior data of a user;
taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through the training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
In one possible design, analyzing a purchase intention value of the user at a future stage through a target learning model according to the online behavior data includes:
according to the online behavior data, analyzing purchase intention values in each future time period through the target learning model respectively; wherein the purchase intention value is used for representing the possibility of the user to purchase the goods in the corresponding time period.
In one possible design, a hidden markov model is constructed according to the purchase intention value of the user in a future stage, and the hidden markov model comprises the following steps:
constructing a state transition matrix according to the purchase intention value of the user at a future stage; the state transition matrix is used for representing the possibility of mutual conversion of purchase intention values in each period;
constructing an observation matrix according to the purchase intention value of the user at a future stage and the historical login behavior of the user; the observation matrix is used for representing the future login condition of the user under different purchase intention values;
setting an initial state vector;
and constructing a hidden Markov model through the state transition matrix, the observation matrix and the initial state vector.
In a second aspect, an embodiment of the present invention provides an apparatus for analyzing a user login behavior, including:
the acquisition module is used for acquiring online behavior data of the user within a preset time period;
the first analysis module is used for analyzing the purchase intention value of the user in the future stage through a target learning model according to the online behavior data;
the building module is used for building a hidden Markov model according to the purchase intention value of the user at a future stage;
and the second analysis module is used for analyzing the login condition of the user at a future stage through the hidden Markov model.
In one possible design, the inline behavior data includes: the method comprises the following steps of commodity browsing record, ordering record, change record of commodities in a shopping cart and attention commodity record.
In one possible design, further comprising: a training module to:
constructing a training sample through historical online behavior data of a user;
taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through the training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
In one possible design, the first analysis module is specifically configured to:
according to the online behavior data, analyzing purchase intention values in each future time period through the target learning model respectively; wherein the purchase intention value is used for representing the possibility of the user to purchase the goods in the corresponding time period.
In one possible design, the building block is specifically configured to:
constructing a state transition matrix according to the purchase intention value of the user at a future stage; the state transition matrix is used for representing the possibility of mutual conversion of purchase intention values in each period;
constructing an observation matrix according to the purchase intention value of the user at a future stage and the historical login behavior of the user; the observation matrix is used for representing the future login condition of the user under different purchase intention values;
setting an initial state vector;
and constructing a hidden Markov model through the state transition matrix, the observation matrix and the initial state vector.
In a third aspect, an embodiment of the present invention provides an analysis system for a user login behavior, including: the device comprises a memory and a processor, wherein the memory stores executable instructions of the processor; wherein the processor is configured to perform the method of analyzing user login behavior of any one of the first aspect via execution of the executable instructions.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for analyzing user login behavior according to any one of the first aspect.
In a fifth aspect, an embodiment of the present invention provides a program product, where the program product includes: a computer program stored in a readable storage medium, the computer program being readable from the readable storage medium by at least one processor of a server, execution of the computer program by the at least one processor causing the server to perform the method of analyzing user login behavior of any of the first aspects.
The invention provides a method, a device and a system for analyzing user login behaviors and a storage medium, which are used for acquiring online behavior data of a user in a preset time period; analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data; constructing a hidden Markov model according to the purchase willingness value of the user in the future stage; and analyzing the login condition of the user in the future stage through the hidden Markov model. The method and the device can analyze the login condition of the future stage according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the future stage of the user, the conversion effect of the recommendation strategy is improved, and the user experience is good.
Drawings
FIG. 1 is a schematic diagram of an application scenario of the present invention;
fig. 2 is a flowchart of a method for analyzing a user login behavior according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a state transition matrix according to an embodiment of the present invention;
fig. 4 is a flowchart of a method for analyzing a user login behavior according to a second embodiment of the present invention;
fig. 5 is a schematic structural diagram of an analysis apparatus for user login behavior according to a third embodiment of the present invention;
fig. 6 is a schematic structural diagram of an apparatus for analyzing a user login behavior according to a fourth embodiment of the present invention;
fig. 7 is a schematic structural diagram of an analysis system of user login behavior according to a fifth embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims, as well as in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are, for example, capable of operation in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
To facilitate understanding of the technical solution, the terms appearing in the present invention are explained.
1) Hidden Markov Models (HMMs) are statistical models that are considered to be a Markov process with unobserved (Hidden) states in the system being modeled. It is used to describe a markov process with hidden unknown parameters. The difficulty is to determine the implicit parameters of the process from the observable parameters. These parameters are then used for further analysis, such as pattern recognition.
2) Random Forest (Random Forest) refers to an algorithm that trains and analyzes samples using multiple decision trees. That is, the random forest algorithm is an algorithm comprising a plurality of decision trees, and the output categories of the random forest algorithm are determined by the numerous trees of the categories output by the individual decision trees. In the sklern module library, functions related to random forest algorithms are all located in an integrated algorithm module ensemble, and the related algorithm functions include a random forest algorithm (random forest classifier), a bagging algorithm (bagging classifier), a complete random tree algorithm (extra tree classifier), an iterative algorithm (adaptive), a GBT gradient Boosting classifier, a gradient regression algorithm (gradientboosting regressor), and a voting algorithm (Votingclassifier).
3) The state transition matrix was proposed by the russian mathematician markov, who discovered in the early 20 th century: some factors of a system are that during the transition, the nth result is only affected by the (n-1) th result, i.e. is only related to the current state, but not to the past state. In markov analysis, the concept of state transition is introduced. The state refers to a state in which an objective thing may appear or exist; state transition refers to the transition of an objective thing from one state to another.
The technical solution of the present invention will be described in detail below with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
More and more user data on the internet are recorded, and the action data of searching, clicking, purchasing and the like of the user on a website reflect the interest and the demand of the user on products. By analyzing and modeling a series of behavior data of the user, the degree of interest of the user in products and websites can be analyzed, so that data support is provided for user recommendation, website popularization and the like. At present, a deep learning model is generally combined to analyze massive online data of a user, and then commodities which the user may order and purchase are analyzed. However, this method can only analyze the purchasing behavior of the logged-in user, and it is difficult to analyze the logging-in behavior of the user at a future stage, so that the user who may log in cannot be recommended in advance, and the recommendation policy effect is poor.
The user login behavior of the e-commerce has close relation with the purchase desire of the user, and the purchase desire is expressed in shopping behaviors of browsing commodities, ordering the commodities, paying attention to the commodities, buying the commodities additionally, reducing the shopping and the like of the user. A supervised learning model can be constructed for the shopping behaviors, purchasing intention values of the user are obtained through a random forest algorithm, a hidden state is defined according to the purchasing intention values, and a hidden state sequence is obtained along with the time; the act of a user logging in may be considered an observable state. Such a pattern of behavior follows a typical hidden markov process. For example, a user purchases some daily necessities, logs in to an e-commerce for several days, and conducts actions of browsing, paying attention, buying additionally, buying negatively, ordering and the like; if the user wants to buy a mobile notebook, the price is about 1 ten thousand, the user can be very careful, the number of times that the user logs in to the e-commerce lasts about one month or more, but the login frequency is probably not very high.
In view of the above technical problems, the present invention provides an analysis method, an apparatus, a system and a storage medium for user login behavior, which can analyze the login situation of a future stage according to the online behavior data of the user, so as to accurately recommend to the user according to the login situation of the future stage of the user, improve the conversion effect of the recommendation strategy and achieve good user experience.
Fig. 1 is a schematic diagram of an application scenario of the present invention, and as shown in fig. 1, the whole diagram is divided into two parts, one part is user behavior and the other part is hidden markov model. The hidden Markov model is divided into two rows of sequences, the first row is an X sequence and the second row is a Y sequence. Each y has an x pointing to it, each y also has another y pointing to it, and each x also has a set of user actions pointing to it.
The definitions are given below:
state sequence: randomly generated state sequences of hidden Markov chains, called state sequences
Observation sequence: each state generates an observation, and the random sequence of observations that results therefrom is called an observation sequence (observation sequence)
Markov model: the Markov model is a probability model about time sequence, and describes a process of generating a random sequence of non-observable states by a hidden Markov chain randomly and generating an observation to generate an observation random sequence by each state. The specific definition is as follows:
let Q be the set of all possible states, and the state of each Q represents the value of the user's will to buy in the next day, and the specific value range is [0.0,1.0] interval. For example, a value of the user's will of purchase at tomorrow is defined to be greater than 0.9, that is, a value in the interval of 0.9 to 1.0 is defined as q 1.
And if the user purchases the wisdom in the next day, the value mapping relation is shown in table 1, and then Q is equal to Q1, Q2, Q3, Q4, Q5, Q6, Q7, Q8, Q9 and Q10.
TABLE 1
Purchase intention value | Status of state | Remarks for note |
0.9~1.0 | q1 | Strong purchase will |
0.8~0.9 | q2 | Willingness to purchase |
0.7~0.8 | q3 | Willingness to purchase |
0.6~0.7 | q4 | Willingness to purchase |
0.5~0.6 | q5 | Willingness to purchase |
0.4~0.5 | q6 | Willingness to purchase |
0.3~0.4 | q7 | Willingness to purchase |
0.2~0.3 | q8 | Willingness to purchase |
0.1~0.2 | q9 | Idea of purchaseWish to |
0.0~0.1 | q10 | The purchase will be extremely weak |
Setting V as the set of all possible observations, wherein the state of each V represents the future day to log in the observation result, the current day is logged in 1, and the current day is not logged in 0, for example, 1 represents the tomorrow login, and is called V1; for example, 01 represents no login tomorrow and a login is acquired, which is called v 2.
The specific mapping relationship of the user login behaviors is shown in table 2, and then V-V1, V2, V3, V4, V5, V6, V7 and V8.
TABLE 2
Wherein Q has 10 state numbers and V has 8 observation numbers.
I is the state sequence of length T and O is the corresponding observation sequence.
I=(i1,i2,...,iT),O=(o1,o2,...,oT)。
A is the state transition matrix: a ═ aij ] nXn, where i ═ 1,2,. 10; j ═ 1, 2.. 10.
Wherein, at time t, the probability of transition to state qj at time t +1 under the condition of qi state: aij is P (it +1 is qj | it is qi).
B is the observation matrix: b ═ bj (k) ] nXm where k ═ 1,2,. 10; j is 1, 2.
Wherein the probability of observing vk is generated under the condition that the time t is in the state qj: bj (k) P (ot) vk | it qj).
π is the initial state vector: pi (pi i), where pi i P (i1 qi).
The hidden Markov model is determined by an initial state vector pi, a state transition matrix A and an observation matrix B. π and A determine the state sequence, B determines the observation sequence. Thus, hidden markov models λ can be represented by ternary symbols, namely: λ ═ a, B, pi. A, B, pi are called three elements of a hidden Markov model.
Analyzing the user login behavior, i.e. the probabilistic calculation problem of the hidden markov model, given a model λ ═ (a, B, pi) and an observation sequence O ═ O1, O2.
By the method, the login condition of the future stage can be analyzed according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the user at the future stage, the conversion effect of the recommendation strategy is improved, and the user experience is good.
The following describes the technical solutions of the present invention and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.
Fig. 2 is a flowchart of a method for analyzing a user login behavior according to an embodiment of the present invention, and as shown in fig. 2, the method in this embodiment may include:
s101, acquiring online behavior data of a user in a preset time period.
In this embodiment, the historical behavior data of the user is the basis of the analysis of the purchase intention and the user login behavior. Therefore, the online behavior data of the user in the preset time period is firstly acquired. Wherein the online behavior data comprises: the method comprises the following steps of commodity browsing record, ordering record, change record of commodities in a shopping cart and attention commodity record. These online behavior data indicate which items the user is interested in, for example, if the user has recently browsed a variety of different types of mobile phones and adds some mobile phones to a shopping cart, it indicates that the user will have a strong desire to purchase the mobile phone at some future time. Meanwhile, the user can be analyzed according to the method, and the possibility of logging in the user in a future period is high. The preset time period can be set to different time lengths according to actual conditions, such as the last week, the last month and the like.
Specifically, first, user shopping behavior data is collected (-6 days, 6 last day). And a big data platform is adopted, shopping behaviors of a user of a supplier for browsing commodities, adding a shopping cart, subtracting the shopping cart, paying attention to the commodities and placing orders are acquired, and the model is modeled in a Hive table form. Specifically modeling: the method comprises the steps that a user account number, a commodity SKU number, a brand ID, a commodity first-level classification ID, a commodity second-level classification ID and a commodity third-level classification ID are measured from the user and the user to the commodity and the commodity. For example, the browsing times in the last 1,2, 3, 5, 7, 15, 30 days, browsing behavior days in the last 7, 15, 30 days, browsing average times in the last 7, 15, 30 days, shopping cart times in the last 1,2, 3, 5, 7, 15, 30 days, shopping cart adding times in the last 7, 15, 30 days, shopping cart average times in the last 1,2, 3, 5, 7, 15, 30 days, shopping cart subtracting times in the last 7, 15, 30 days, attention times in the last 1,2, 3, 5, 7, 15, 30 days, attention behavior days in the last 7, 15, 30 days, attention average times in the last 1,2, 3, 5, 7, 15, 30 days, and order taking times in the last 7, 15, 30 days, order taking times in the last 7, 30 days, the single average times in the last 7, 15 and 30 days, the action times in the last 1,2, 3, 5, 7, 15 and 30 days are weighted and summed, and the conversion rate in the last 1,2, 3, 5, 7, 15 and 30 days is obtained. A total of 200 multidimensional data. Then, single data were collected for approximately 5 days. Under a big data platform, acquiring order data of a user: the user account number, the commodity SKU number, the brand ID, the commodity primary classification ID, the commodity secondary classification ID, the commodity tertiary classification ID and the order placing quantity are stored in the Hive table. The ordering data is then used to mark shopping behavior data. And marking shopping behavior data of the user by adopting ordering data of nearly 5 days, namely, correlating two hive tables, wherein for the behavior data with ordering, the marking Label value is 1, and if not, the marking Label value is 0. And finally forming marked user shopping behavior data: and storing the user behavior data + the marking result Label (1 or 0) into a Hive table, namely a user shopping behavior data table for ordering. And finally, collecting the shopping behavior data of the user today. And collecting the shopping behavior data of the user on the current day, namely the shopping behavior data of the 200 multidimensional user, and storing the data into the Hive table under a big data platform.
And S102, analyzing the purchase intention value of the user in the future stage through a target learning model according to the online behavior data.
In the embodiment, according to the online behavior data, the purchase intention values in each future time period are respectively analyzed through the target learning model; wherein, the purchase intention value is used for representing the possibility that the user purchases the goods in the corresponding time period.
Specifically, a random forest algorithm trains marked shopping behavior data of the user and analyzes the purchasing intention of the user at each future stage. Firstly, establishing a Spark task on a big data platform, wherein the specific small steps comprise: using Spark SQL to inquire the marked user behavior data prepared in the front; classifying the marked user behavior data by using a random forest classifier class of Spark to generate a random forest classification model; using a random forest classification model to analyze user behavior data of the same day, generating a user tomorrow purchase intention value, storing the user purchase intention value into a user purchase intention analysis Hive table, wherein the user purchase intention Hive table user _ buy _ prob can represent that: user _ log _ acct (user account), buy _ prob (purchase will), dt (date partition). Then, the user's will of purchase in the next day for 30 days may also be calculated. Repeating the step 29 times, obtaining the user's own will for 30 days by adding the first time, and storing the value of the user's own will for 30 days into the user's purchase will analysis Hive table. Therefore, the tomorrow purchase intention value of the user of the e-commerce for 30 days is calculated, for example, the user log acct (user account) is queried as the 30-day purchase intention value of "jack", and the Hive table user _ buy _ prob can be queried and received through SQL statements, namely, select buy _ prob and dt from user _ buy _ prob. After the big data platform is executed, 6-day user purchase intention records of the user Jack are displayed as shown in table 3, namely the user Jack purchase intention state changes q1- > q2- > q9- > q10- > q7- > q3 in the time period of 2018-09-01-2018-09-06.
TABLE 3
Date | Specific will value | Corresponding state |
2018-09-01 | 0.91 | q1 |
2018-09-02 | 0.81 | q2 |
2018-09-03 | 0.13 | q9 |
2018-09-04 | 0.01 | q10 |
2018-09-05 | 0.32 | q7 |
2018-09-06 | 0.78 | q3 |
It should be noted that, this embodiment does not limit the random forest classification algorithm, and theoretically, other classification algorithms may also be applicable to the present invention, and those skilled in the art may select other learning models according to actual situations.
S103, constructing a hidden Markov model according to the purchase intention value of the user in the future stage.
In this embodiment, constructing a hidden markov model according to a purchase intention value of a user at a future stage includes: constructing a state transition matrix according to a purchase intention value of a user at a future stage; the state transition matrix is used for representing the possibility of mutual conversion of purchase intention values in each time period; constructing an observation matrix according to a purchase intention value of a user at a future stage and historical login behaviors of the user; the observation matrix is used for representing the future login condition of the user under different purchase intention values; setting an initial state vector; and constructing a hidden Markov model through the state transition matrix, the observation matrix and the initial state vector.
Specifically, according to the purchase intention value of the user in the future stage, a hidden Markov model is constructed, which comprises the following steps:
and (I) calculating a state transition matrix A.
On a large data platform, a 30-day user purchase desire value record of the hive table user _ buy _ prob (user _ log _ acct, buy _ prob, dt) is obtained through the previous calculation. Fig. 3 is a schematic diagram of a principle of a state transition matrix according to an embodiment of the present invention, and as shown in fig. 3, q1- > q1, q1- > q2, q1- > q3, …, q1- > q10 can also be expressed as a11, a12, a13, … a110 of an a matrix. The first a11 is obtained, and the other matrix elements are calculated in the same manner. A11 calculation step:
the set USER _ a of the USER _ log _ acct satisfying the condition that dt is '2018-09-01' and buy _ prob > -0.9 is represented by SQL, i.e., select USER _ log _ access from USER _ buy _ prob where dt is '2018-09-01' and buy _ prob > -0.9.
The set USER _ B of the USER _ log _ acct satisfying the condition that dt is '2018-09-02' and buy _ prob > -0.9 is represented as SQL, and the select USER _ log _ access from USER _ buy _ prob where dt is '2018-09-02' and buy _ prob > -0.9.
The set USER _ C ═ USER _ a ≈ USER _ B means that the USER whose purchase intention state is q1 is converted from the USER set of USERs whose date 2018-09-02 purchase intention state is q1, by the expression q1- > q 1.
The number of USER _ C/number of USER _ a, a11, is, for example, 0.67.
By analogy, a 12-0.03, a 13-0.05, a 14-0.05, a 15-0.05, a 16-0.05, a 17-0.03, a 18-0.02, a 19-0.03, and a 110-0.02 are calculated.
Similarly, q2, q3, q4, and q … q10 can be calculated to obtain other state transition values, and the state transition matrix a is finally:
(II) calculating an observation matrix B
On a big data platform, a user login stream record of an e-commerce is stored in a Hive table user _ logic, and a specific field is as follows: user _ log _ acct (user account number), logic _ num (login number), dt (date partition). The probability of observing vk is determined by determining bj (k) P (ot) vk | it qj) and setting the state qj at time t.
First, b1(1) ═ P (ot ═ v1| it ═ q1) is obtained, and other matrix elements are calculated in the same way:
the set USER _ Q1 of the USER _ log _ acc satisfying the condition that dt is '2018-09-01' and buy _ prob > is 0.9 is expressed as SQL, and select USER _ log _ acc from USER _ buy _ prob where dt is '2018-09-01' and buy _ prob > is 0.9.
The set USER _ V1 of USER _ logic satisfying the condition dt ═ 2018-09-01'and logic _ num > -1 is expressed as SQL, i.e. select USER _ log _ acc from USER _ logic where dt ═ 2018-09-01' and logic _ num > -1.
And (4) obtaining an intersection USER _ Q1V1 ═ USER _ Q1 ═ USER _ V1, namely that the tomorrow purchase intention of the USER is more than 0.9 and the USER logs in tomorrow.
b1(1) ═ number of sets of USER _ Q1V 1/number of sets of USER _ Q such as: b1(1) is 0.71.
Similarly calculate b1(2):
the set USER _ V2 of USER _ logic satisfying the condition dt ═ 2018-09-02'and logic _ num > -1 is expressed as SQL, i.e. select USER _ log _ acc from USER _ logic where dt ═ 2018-09-01' and logic _ num > -1.
The REAL _ USER _ V2 is a set obtained by subtracting tomorrow login from the acquired login set of USER _ V2-USER _ V1, that is, a set of USERs who have not logged in tomorrow and have logged in tomorrow.
Intersection is obtained (namely the tomorrow purchase intention of the USER is more than 0.9, and the tomorrow is not logged in and is logged in later days), the USER _ Q1V2 is equal to USER _ Q1 and equal to REAL _ USER _ V2
b1(2) ═ number of sets of USER _ Q1V 2/number of sets of USER _ Q such as: b1(2) is 0.09.
By analogy, b1(3) is 0.03, b1(4) is 0.05, b1(5) is 0.02, b1(6) is 0.05, b1(7) is 0.01, b1(8) is 0.04
Similarly, the conversion values from q2, q3, q4, … q10 to the observation state can be calculated, and finally the observation matrix B is obtained as:
(III) constructing a hidden Markov model
The A, B matrices have been previously calculated to give a pi of, say, (1,0,0,0,0,0, 0)TThe hidden markov model λ ═ (a, B, pi) is constructed.
And S104, analyzing the login condition of the user in the future stage through a hidden Markov model.
In the embodiment, the login condition of the user at a future stage is analyzed through the hidden Markov model.
Specifically, the hidden markov model has been constructed in the previous step, and it can be analyzed by the hidden markov model that a certain user has a higher probability of logging in on a certain day. For example, given user Jack, the initial Q state is Q1, indicating a strong purchase intention, and pi is (1,0,0,0,0, 0)TObservation sequence O ═ { v1, v1} analysis user Jack tomorrow afterday loginProbability P (O | λ).
Probability of user Jack logging in tomorrow:
a1(1)=π1b1(O1)=1*0.71=0.71
a1(2)=π2b2(O1)=0*0.61=0
all the remaining values can be calculated by analogy;
for example: a1(10) ═ pi 10b10(O1) ═ 0.04 ═ 0
The probability of the user logging in tomorrow is 0.71.
Similarly, the probability of the Jack logging in the next day can be calculated in the same way.
In the embodiment, online behavior data of a user in a preset time period is acquired; analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data; constructing a hidden Markov model according to a purchase intention value of a user at a future stage; and analyzing the login condition of the user in a future stage through a hidden Markov model. The method and the device can analyze the login condition of the future stage according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the future stage of the user, the conversion effect of the recommendation strategy is improved, and the user experience is good.
Fig. 4 is a flowchart of an analysis method for user login behavior according to a second embodiment of the present invention, and as shown in fig. 4, the method in this embodiment may include:
s201, constructing a training sample, and obtaining a target learning model through iteration of the training sample.
In the embodiment, a training sample is constructed through historical online behavior data of a user; taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through a training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model. The historical online behavior data of the user can be obtained according to S101, the purchase condition of the user in the future stage is output by the learning model, and the value range is a numerical value between 0 and 1; for example, if the purchase probability is greater than 50%, and the corresponding real ordering record is successful, the analysis result is considered to be correct. The construction method of the learning model is the prior art and is not described herein again. It should be noted that the random forest classification algorithm is adopted in the embodiment, the learning model is not limited to the random forest classification algorithm, and other classification algorithms can also be applied to the invention in theory.
S202, acquiring online behavior data of the user in a preset time period.
And S203, analyzing the purchase intention value of the user in the future stage through the target learning model according to the online behavior data.
S204, constructing a hidden Markov model according to the purchase intention value of the user in the future stage.
S205, analyzing the login situation of the user in the future stage through a hidden Markov model.
In this embodiment, please refer to the relevant description in step S101 to step S104 in the method shown in fig. 2 for the specific implementation process and technical principle of step S202 to step S205, which is not described herein again.
In the embodiment, online behavior data of a user in a preset time period is acquired; analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data; constructing a hidden Markov model according to a purchase intention value of a user at a future stage; and analyzing the login condition of the user in a future stage through a hidden Markov model. The method and the device can analyze the login condition of the future stage according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the future stage of the user, the conversion effect of the recommendation strategy is improved, and the user experience is good.
In addition, the embodiment can also construct training samples through historical online behavior data of the user; taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through a training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
Fig. 5 is a schematic structural diagram of an analysis apparatus for a user login behavior according to a third embodiment of the present invention, and as shown in fig. 5, the analysis apparatus for a user login behavior according to the present embodiment may include:
the acquiring module 31 is configured to acquire online behavior data of a user within a preset time period;
the first analysis module 32 is used for analyzing the purchase intention value of the user in the future stage through the target learning model according to the online behavior data;
the building module 33 is used for building a hidden Markov model according to the purchase intention value of the user at a future stage;
and the second analysis module 34 is used for analyzing the login situation of the user at a future stage through the hidden Markov model.
In one possible design, the on-line behavior data includes: the method comprises the following steps of commodity browsing record, ordering record, change record of commodities in a shopping cart and attention commodity record.
In one possible design, the first analysis module 32 is specifically configured to:
according to the online behavior data, analyzing purchase intention values in each future time period through a target learning model; wherein, the purchase intention value is used for representing the possibility that the user purchases the goods in the corresponding time period.
In one possible design, the building block 33 is specifically configured to:
constructing a state transition matrix according to a purchase intention value of a user at a future stage; the state transition matrix is used for representing the possibility of mutual conversion of purchase intention values in each time period;
constructing an observation matrix according to a purchase intention value of a user at a future stage and historical login behaviors of the user; the observation matrix is used for representing the future login condition of the user under different purchase intention values;
setting an initial state vector;
and constructing a hidden Markov model through the state transition matrix, the observation matrix and the initial state vector.
The analysis apparatus for user login behavior in this embodiment may execute the technical solution in the method shown in fig. 2, and for specific implementation processes and technical principles, reference is made to the relevant description in the method shown in fig. 2, which is not described herein again.
In the embodiment, online behavior data of a user in a preset time period is acquired; analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data; constructing a hidden Markov model according to a purchase intention value of a user at a future stage; and analyzing the login condition of the user in a future stage through a hidden Markov model. The method and the device can analyze the login condition of the future stage according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the future stage of the user, the conversion effect of the recommendation strategy is improved, and the user experience is good.
Fig. 6 is a schematic structural diagram of an analysis apparatus for a user login behavior according to a fourth embodiment of the present invention, as shown in fig. 6, the analysis apparatus for a user login behavior according to the present embodiment may further include, on the basis of the apparatus shown in fig. 5:
the training module 35 is specifically configured to:
constructing a training sample through historical online behavior data of a user;
taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through a training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
The analysis apparatus for user login behavior in this embodiment may execute the technical solutions in the methods shown in fig. 2 and fig. 4, and specific implementation processes and technical principles of the analysis apparatus refer to the relevant descriptions in the methods shown in fig. 2 and fig. 4, which are not described herein again.
In the embodiment, online behavior data of a user in a preset time period is acquired; analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data; constructing a hidden Markov model according to a purchase intention value of a user at a future stage; and analyzing the login condition of the user in a future stage through a hidden Markov model. The method and the device can analyze the login condition of the future stage according to the online behavior data of the user, so that accurate recommendation can be performed on the user according to the login condition of the future stage of the user, the conversion effect of the recommendation strategy is improved, and the user experience is good.
In addition, the embodiment can also construct training samples through historical online behavior data of the user; taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through a training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
Fig. 7 is a schematic structural diagram of an analysis system of a user login behavior according to a fifth embodiment of the present invention, and as shown in fig. 7, the analysis system 40 of a user login behavior according to this embodiment may include: a processor 41 and a memory 42.
A memory 42 for storing programs; the Memory 42 may include a volatile Memory (RAM), such as a Static Random Access Memory (SRAM), a Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM), and the like; the memory may also comprise a non-volatile memory, such as a flash memory. The memory 42 is used to store computer programs (e.g., applications, functional modules, etc. that implement the above-described methods), computer instructions, etc., which may be stored in one or more of the memories 42 in a partitioned manner. And the above-mentioned computer program, computer instructions, data, etc. can be called by the processor 41.
The computer programs, computer instructions, etc. described above may be stored in one or more memories 42 in partitions. And the above-mentioned computer program, computer instructions, data, etc. can be called by the processor 41.
A processor 41 for executing the computer program stored in the memory 42 to implement the steps of the method according to the above embodiments.
Reference may be made in particular to the description relating to the preceding method embodiment.
The processor 41 and the memory 42 may be separate structures or may be integrated structures integrated together. When the processor 41 and the memory 42 are separate structures, the memory 42 and the processor 41 may be coupled by a bus 43.
The analysis system for the user login behavior in this embodiment may execute the technical solutions in the methods shown in fig. 2 and fig. 4, and specific implementation processes and technical principles thereof refer to the relevant descriptions in the methods shown in fig. 2 and fig. 4, which are not described herein again.
In addition, embodiments of the present application further provide a computer-readable storage medium, in which computer-executable instructions are stored, and when at least one processor of the user equipment executes the computer-executable instructions, the user equipment performs the above-mentioned various possible methods.
Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a general purpose or special purpose computer. An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. Of course, the storage medium may also be integral to the processor. The processor and the storage medium may reside in an ASIC. Additionally, the ASIC may reside in user equipment. Of course, the processor and the storage medium may reside as discrete components in a communication device.
The present application further provides a program product comprising a computer program stored in a readable storage medium, from which the computer program can be read by at least one processor of a server, the execution of the computer program by the at least one processor causing the server to carry out the method of any of the embodiments of the invention described above.
Those of ordinary skill in the art will understand that: all or a portion of the steps of implementing the above-described method embodiments may be performed by hardware associated with program instructions. The program may be stored in a computer-readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for analyzing user login behavior is characterized by comprising the following steps:
acquiring online behavior data of a user within a preset time period;
analyzing a purchase intention value of the user in a future stage through a target learning model according to the online behavior data;
constructing a hidden Markov model according to the purchase willingness value of the user in the future stage;
and analyzing the login condition of the user in the future stage through the hidden Markov model.
2. The method of claim 1, wherein the online behavior data comprises: the method comprises the following steps of commodity browsing record, ordering record, change record of commodities in a shopping cart and attention commodity record.
3. The method of claim 1, further comprising:
constructing a training sample through historical online behavior data of a user;
taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through the training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
4. The method of claim 1, wherein analyzing a purchase intent value of the user at a future stage according to the online behavior data through a target learning model comprises:
according to the online behavior data, analyzing purchase intention values in each future time period through the target learning model respectively; wherein the purchase intention value is used for representing the possibility of the user to purchase the goods in the corresponding time period.
5. The method of claim 4, wherein constructing a hidden Markov model based on the user's willingness to purchase value at a future stage comprises:
constructing a state transition matrix according to the purchase intention value of the user at a future stage; the state transition matrix is used for representing the possibility of mutual conversion of purchase intention values in each period;
constructing an observation matrix according to the purchase intention value of the user at a future stage and the historical login behavior of the user; the observation matrix is used for representing the future login condition of the user under different purchase intention values;
setting an initial state vector;
and constructing a hidden Markov model through the state transition matrix, the observation matrix and the initial state vector.
6. An apparatus for analyzing a user login behavior, comprising:
the acquisition module is used for acquiring online behavior data of the user within a preset time period;
the first analysis module is used for analyzing the purchase intention value of the user in the future stage through a target learning model according to the online behavior data;
the building module is used for building a hidden Markov model according to the purchase intention value of the user at a future stage;
and the second analysis module is used for analyzing the login condition of the user at a future stage through the hidden Markov model.
7. The apparatus of claim 6, wherein the inline behavior data comprises: the method comprises the following steps of commodity browsing record, ordering record, change record of commodities in a shopping cart and attention commodity record.
8. The apparatus of claim 6, further comprising: a training module to:
constructing a training sample through historical online behavior data of a user;
taking a real ordering record as a target output result, and iteratively training the constructed initial learning model through the training sample until the error between the analysis result output by the trained initial learning model and the real ordering record is smaller than a preset threshold value; and obtaining the target learning model.
9. An analysis system for user login behavior, comprising: the memory is used for storing executable instructions of the processor; wherein the processor is configured to perform the method of analyzing user login behavior of claims 1-5 via execution of the executable instructions.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method for analyzing a user login behavior according to any one of claims 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910636193.2A CN112231548A (en) | 2019-07-15 | 2019-07-15 | User login behavior analysis method, device and system and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910636193.2A CN112231548A (en) | 2019-07-15 | 2019-07-15 | User login behavior analysis method, device and system and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112231548A true CN112231548A (en) | 2021-01-15 |
Family
ID=74111456
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910636193.2A Pending CN112231548A (en) | 2019-07-15 | 2019-07-15 | User login behavior analysis method, device and system and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112231548A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113592551A (en) * | 2021-07-31 | 2021-11-02 | 广州小鹏汽车科技有限公司 | Method, device and equipment for analyzing and processing behavior data of vehicle purchasing user |
CN113793180A (en) * | 2021-09-15 | 2021-12-14 | 北京沃东天骏信息技术有限公司 | User preference analysis method, device, equipment and computer storage medium |
CN117557346A (en) * | 2024-01-11 | 2024-02-13 | 华高数字科技有限公司 | Full-link intelligent business decision analysis method based on dynamic consumption data |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009093512A (en) * | 2007-10-11 | 2009-04-30 | Nec Corp | Log-in information processing system and log-in information processing method |
JP2014160345A (en) * | 2013-02-19 | 2014-09-04 | Nippon Telegr & Teleph Corp <Ntt> | Browsing action predicting device, browsing action learning device, browsing action predicting method, and browsing action learning method and program |
CN104978425A (en) * | 2015-07-02 | 2015-10-14 | 广东小天才科技有限公司 | Method and apparatus for pushing learning contents for user |
CN108491267A (en) * | 2018-03-13 | 2018-09-04 | 百度在线网络技术(北京)有限公司 | Method and apparatus for generating information |
CN109934704A (en) * | 2019-03-22 | 2019-06-25 | 深圳乐信软件技术有限公司 | Information recommendation method, device, equipment and storage medium |
-
2019
- 2019-07-15 CN CN201910636193.2A patent/CN112231548A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009093512A (en) * | 2007-10-11 | 2009-04-30 | Nec Corp | Log-in information processing system and log-in information processing method |
JP2014160345A (en) * | 2013-02-19 | 2014-09-04 | Nippon Telegr & Teleph Corp <Ntt> | Browsing action predicting device, browsing action learning device, browsing action predicting method, and browsing action learning method and program |
CN104978425A (en) * | 2015-07-02 | 2015-10-14 | 广东小天才科技有限公司 | Method and apparatus for pushing learning contents for user |
CN108491267A (en) * | 2018-03-13 | 2018-09-04 | 百度在线网络技术(北京)有限公司 | Method and apparatus for generating information |
CN109934704A (en) * | 2019-03-22 | 2019-06-25 | 深圳乐信软件技术有限公司 | Information recommendation method, device, equipment and storage medium |
Non-Patent Citations (1)
Title |
---|
张玉成;徐大纹;王筱娟;: "基于加权马尔可夫链的主动用户行为预测模型", 计算机工程与设计, no. 10, 16 October 2011 (2011-10-16), pages 3334 - 3337 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113592551A (en) * | 2021-07-31 | 2021-11-02 | 广州小鹏汽车科技有限公司 | Method, device and equipment for analyzing and processing behavior data of vehicle purchasing user |
CN113793180A (en) * | 2021-09-15 | 2021-12-14 | 北京沃东天骏信息技术有限公司 | User preference analysis method, device, equipment and computer storage medium |
CN117557346A (en) * | 2024-01-11 | 2024-02-13 | 华高数字科技有限公司 | Full-link intelligent business decision analysis method based on dynamic consumption data |
CN117557346B (en) * | 2024-01-11 | 2024-04-02 | 华高数字科技有限公司 | Full-link intelligent business decision analysis method based on dynamic consumption data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108898459B (en) | Commodity recommendation method and device | |
CN107341716B (en) | Malicious order identification method and device and electronic equipment | |
CN108985830B (en) | Recommendation scoring method and device based on heterogeneous information network | |
Shastri et al. | Stock price prediction using artificial neural model: an application of big data | |
CN106251174A (en) | Information recommendation method and device | |
Chatterjee et al. | Stock price prediction using time series, econometric, machine learning, and deep learning models | |
CN112231548A (en) | User login behavior analysis method, device and system and storage medium | |
Basallo-Triana et al. | Analogue-based demand forecasting of short life-cycle products: a regression approach and a comprehensive assessment | |
CN111080360A (en) | Behavior prediction method, model training method, device, server and storage medium | |
CN113191838B (en) | Shopping recommendation method and system based on heterogeneous graph neural network | |
CN111695024A (en) | Object evaluation value prediction method and system, and recommendation method and system | |
Goyal et al. | A survey on ensemble model for loan prediction | |
CN107292713A (en) | A kind of rule-based individual character merged with level recommends method | |
CN116362823A (en) | Recommendation model training method, recommendation method and recommendation device for behavior sparse scene | |
Ramachandra et al. | Machine learning application for black friday sales prediction framework | |
Proietti et al. | Euromind‐: A Density Estimate of Monthly Gross Domestic Product for the Euro Area | |
Winata et al. | Predicting stock market prices using time series sarima | |
Ye et al. | Forecasting seasonal demand for retail: A Fourier time-varying grey model | |
Huynh et al. | Causal inference in econometrics | |
Kanade et al. | Machine learning model for stock market prediction | |
CN110766465A (en) | Financial product evaluation method and verification method and device thereof | |
JP5835802B2 (en) | Purchase forecasting apparatus, method, and program | |
WO2013055257A1 (en) | Method for predicting a target for events on the basis of an unlimited number of characteristics | |
CN113159927A (en) | Method and device for determining client label | |
CN114529399A (en) | User data processing method, device, computer equipment and storage medium |
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 |