CN107103487B - Advertisement pushing method based on big data - Google Patents
Advertisement pushing method based on big data Download PDFInfo
- Publication number
- CN107103487B CN107103487B CN201710120950.1A CN201710120950A CN107103487B CN 107103487 B CN107103487 B CN 107103487B CN 201710120950 A CN201710120950 A CN 201710120950A CN 107103487 B CN107103487 B CN 107103487B
- Authority
- CN
- China
- Prior art keywords
- commodity
- webpage
- promoted
- classification
- score
- 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.)
- Active
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
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0257—User requested
-
- 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
- G06Q30/0241—Advertisements
- G06Q30/0277—Online advertisement
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Information Transfer Between Computers (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an advertisement pushing method based on big data, which comprises the following steps: according to the big data accumulation, determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user; determining the scores of G commodity classifications with the strongest relevance of all webpage commodities to be promoted, the scores of H webpage type classifications with the strongest relevance of currently accessed webpages, and the correlation weight between each commodity classification and each webpage type classification; determining a total pushing score of each commodity according to a correlation weight between the commodity classification and the webpage type classification, a corresponding commodity classification score, a webpage type classification score and an attention value of the webpage commodity; sorting the commodities according to the total pushed scores of the commodities from high to low; and pushing the first N webpage commodities to be promoted in the sequencing result through N preassigned advertisement positions. The method can push the commodities with high interest probability of the user, and realize accurate marketing.
Description
Technical Field
The invention relates to the technical field of big data analysis and the field of internet advertisement putting, in particular to an advertisement pushing method based on big data.
Background
With the development of the internet, the commodity popularization through the network becomes a new marketing mode. For vast netizens, what is needed is an accurately delivered advertisement, rather than bombing of various types of advertisements, and too many advertisements can bring about a bad web browsing experience, and on the other hand, for advertisement delivering merchants and owners, the accurately delivered advertisement can save advertisement propaganda cost and bring about better commodity sales drainage.
At present, in order to better serve users and more accurately put network advertisements, a system for recording user browsing behaviors and performing simple analysis appears on the internet, and the user behaviors are recorded and analyzed and then guided by advertisements promoted on webpages.
Disclosure of Invention
The invention provides an advertisement pushing method based on big data, which is used for solving the problem that the existing network-pushed advertisement can not meet the user requirements particularly and accurately.
The invention provides an advertisement pushing method based on big data, which comprises the following steps:
according to the big data accumulation, determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user;
according to big data accumulation, determining the scores of G product classifications with strongest relevance in a plurality of product classifications corresponding to each webpage product to be popularized, determining the scores of H webpage classification with strongest relevance in a plurality of webpage classification corresponding to the currently visited webpage, and determining the correlation weight between each product classification in the G product classifications and each webpage classification in the H webpage classification;
for each webpage commodity to be promoted, determining a total commodity pushing score according to a correlation weight value between each commodity classification in the G commodity classifications and each webpage type classification in the H webpage type classifications, and a corresponding commodity classification score, a webpage type classification score and an attention value of the webpage commodity to be promoted;
sequencing a plurality of webpage commodities to be promoted from high to low according to the total score pushed by the commodities;
and pushing the first N webpage commodities to be promoted in the sequencing result through N preassigned advertisement positions.
In one embodiment, the determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user includes:
according to a preset attention value calculation formula, calculating the attention value of each webpage commodity to be promoted according to the commodity access frequency of the webpage commodity to be promoted to the current user, the user residence time and the commodity exposure frequency; the interest value of the commodity to be promoted in the interest value calculation formula is an increasing function of commodity access frequency, an increasing function of user residence time and a decreasing function of commodity exposure frequency.
In one embodiment, the attention value calculation formula is Uv-a × Fv + b × Tv-c × Ev, wherein Uv is the attention value of the vth-th webpage commodity to be promoted, Fv is the commodity access frequency of the current vth-th webpage commodity to be promoted to the current user, Tv is the user residence time of the current vth-th webpage commodity to be promoted to the current user, Ev is the commodity exposure frequency of the current vth-th webpage commodity to be promoted to the current user, a, b and c are pre-specified positive coefficients, v-1, 2.
In one embodiment, a is 1, b is 1, and c is 1.
In one embodiment, the attention value calculation formula is: uv ═ lgFv + lgTv-lgEv; the method comprises the following steps that Uv is an attention value of a vth webpage commodity to be promoted, Fv is commodity access frequency of a current vth webpage commodity to be promoted to a current user, Tv is user residence time of the current vth webpage commodity to be promoted to the current user, and Ev is commodity exposure frequency of the current vth webpage commodity to be promoted to the current user; 1,2,. n; and n is the number of webpage commodities to be promoted.
In one embodiment, the determining, for each webpage commodity to be promoted, a total commodity pushing score according to a correlation weight between each commodity category in the G commodity categories and each webpage category in the H webpage category categories, a score of the corresponding commodity category, a score of the corresponding webpage category, and an attention value of the webpage commodity to be promoted includes:
for each webpage commodity to be promoted, according to the formula Aij＝Uv×Rvi×Pj×WvijCalculating to obtain G × H total scores to be selected of the webpage commodities to be promoted, wherein Uv is the attention value of the vth webpage commodity to be promoted, and RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjFor the currently visited webpage pairThe corresponding value of the jth webpage type classification in the H webpage type classifications; wvijThe correlation weight between the ith commodity classification in the G commodity classifications corresponding to the vth webpage commodity to be promoted and the jth webpage type classification in the H webpage type classifications; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; n is the number of webpage commodities to be promoted;
and taking the highest one of the G × H total scores to be selected of each webpage commodity to be promoted as the commodity pushing total score of the webpage commodity to be promoted.
In one embodiment, the determining, for each webpage commodity to be promoted, a total commodity pushing score according to a correlation weight between each commodity category in the G commodity categories and each webpage category in the H webpage category categories, a score of the corresponding commodity category, a score of the corresponding webpage category, and an attention value of the webpage commodity to be promoted includes:
for each webpage commodity to be promoted, according to a formula:
calculating to obtain G × H total scores to be selected of the webpage commodities to be promoted, wherein Uv is the attention value of the vth webpage commodity to be promoted, and RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjThe score of the jth webpage type classification in the H webpage type classifications corresponding to the currently accessed webpage is obtained; wvijThe correlation weight between the ith commodity classification in the G commodity classifications corresponding to the vth webpage commodity to be promoted and the jth webpage type classification in the H webpage type classifications; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; n is the number of webpage commodities to be promoted;
and taking the highest one of the G × H total scores to be selected of each webpage commodity to be promoted as the commodity pushing total score of the webpage commodity to be promoted.
In an embodiment, for each webpage commodity to be promoted, determining a total commodity pushing score according to a correlation weight between each commodity classification and each webpage type classification corresponding to the webpage commodity to be promoted, and a score of the corresponding commodity classification, a score of the webpage type classification, and an attention value of the webpage commodity to be promoted includes:
for each webpage commodity to be promoted, according to the formula Aij ═ lg (Uv × Rv)i×Pj×Wvij) Calculating to obtain G × H total scores to be selected of the webpage commodities to be promoted, wherein Uv is the attention value of the vth webpage commodity to be promoted, and RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjThe score of the jth webpage type classification in the H webpage type classifications corresponding to the currently accessed webpage is obtained; wvijThe correlation weight between the ith commodity classification in the G commodity classifications corresponding to the vth webpage commodity to be promoted and the jth webpage type classification in the H webpage type classifications; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; n is the number of webpage commodities to be promoted;
and taking the highest one of the G × H total scores to be selected of each webpage commodity to be promoted as the commodity pushing total score of the webpage commodity to be promoted.
In one embodiment, G-3 and H-3.
In one embodiment, after the pushing the top N to-be-promoted webpage commodities in the sorting result through the N pre-specified advertisement slots, the method further includes:
receiving a selection of the webpage commodity pushed in the advertisement space;
accessing the selected webpage commodity and starting timing the browsing time of the user, meanwhile increasing the commodity exposure frequency of the webpage commodity pushed in the advertisement space but not selected to the current user by a first preset step length, and increasing the commodity access frequency of the selected webpage commodity to the current user by a second preset step length; the first preset step length is far smaller than the second preset step length;
receiving an access ending instruction for the selected webpage commodity;
ending the access to the selected web page commodity according to the formula
Updating the user residence time of the selected webpage commodity to the current user; wherein, T is the browsing time length of this time, T is the preset time length, T0Is a third preset step length of]Is a rounded symbol.
In one embodiment, after the pushing the top N to-be-promoted webpage commodities in the sorting result through the N pre-specified advertisement slots, the method further includes:
receiving a selection of the webpage commodity pushed in the advertisement space;
accessing the selected webpage commodity and timing the current browsing time of the user, increasing the exposure frequency of the webpage commodity pushed in the advertisement space but not selected to the current user commodity by a first step length, and increasing the commodity access frequency of the selected webpage commodity to the current user by a second step length;
when the browsing time length is increased by a preset unit time length, increasing a third step length for the user residence time length of the current user by the selected webpage commodity;
receiving an access ending instruction for the selected webpage commodity;
and ending the access to the selected webpage commodity.
In one embodiment, a formula is employed
Calculating a first step length; using a formula
Calculating a second step length; using a formula
Calculating a third step length;
wherein Se is a first step length, kE1Is pre-specifiedExposure growth factor, N, greater than zeroEFor counting the number of exposures, kE2Is a preset exposure power coefficient, and 0 < kE2Less than 1; sf is the second step, kF1For a pre-specified access growth factor, 0 < kE1＜kF1，NFFor the counted number of accesses, kF2For a predetermined access power coefficient, 0 < kE2＜kF2Less than 1, Tv is the user residence time of the selected webpage commodity to the current user, delta TmaxTo preset the maximum dwell time, kTIs a predetermined attenuation factor.
In one embodiment, the initial value of the commodity access frequency of the webpage commodity to be promoted to the current user is 0 or a first preset initial value; the initial value of the user residence time of the webpage commodity to be promoted to the current user is 0 or a second preset initial value; and the initial value of the commodity exposure frequency of the webpage commodity to be promoted to the current user is a third preset initial value.
In one embodiment, after receiving the selection of the webpage goods pushed in the advertisement space, the method further comprises:
improving the commodity classification score, the webpage type classification score and the correlation weight between the commodity classification and the webpage type classification corresponding to the commodity push total score of the selected webpage commodity; and reducing the commodity classification scores, the webpage type classification scores and the correlation weights between the commodity classifications and the webpage type classifications corresponding to the commodity pushing total scores of the webpage commodities pushed but not selected in the advertisement space.
Some of the benefits of the present invention may include:
according to the advertisement pushing method based on the big data, the webpage commodities to be promoted are scored according to the commodity access frequency of the webpage commodities to be promoted to the current user, the residence time of the user, the commodity exposure frequency, the commodity pair classification relevance, the webpage pair classification relevance and the classification relevance of the commodities to be promoted through big data accumulation, the total scores of the pushed commodities are ranked, and the pushing priorities of the commodities to be pushed are determined according to the ranked optimal results. Meanwhile, all factors influencing the total pushing score of the webpage commodity to be pushed are dynamically adjusted according to the selection of the user, so that dynamic self-consistency under big data is achieved. The method can dynamically calculate the commodities with higher interest probability of the user and push the commodities to the user, so that the aim of accurate marketing is fulfilled, and the webpage use experience of the user is improved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flowchart of a big data-based advertisement delivery method according to an embodiment of the present invention;
fig. 2 is a flowchart of another advertisement push method based on big data according to an embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
Fig. 1 is a flowchart of an advertisement push method based on big data according to an embodiment of the present invention. As shown in fig. 1, the method comprises the following steps S101-S105:
s101: and according to the big data accumulation, determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user.
In the embodiment, according to a preset attention value calculation formula, the attention value of each webpage commodity to be promoted is calculated according to the commodity access frequency of the webpage commodity to be promoted to the current user, the residence time of the user and the commodity exposure frequency; the interest value of the commodity to be promoted in the interest value calculation formula is an increasing function of commodity access frequency, an increasing function of user residence time and a decreasing function of commodity exposure frequency. Namely: the commodity access frequency is increased, which means that the commodity is popular with users, and the attention value corresponding to the commodity to be promoted is increased; the residence time of the user is increased, which means that the time for the user to browse the webpage commodity is longer, and shows that the user probably pays attention to the commodity, and the attention value corresponding to the commodity to be promoted is improved; a commodity repeatedly appears on a user page, if the user does not move, the commodity is not interested by the user, and therefore the attention value of the commodity to be promoted is a decreasing function of the exposure of the commodity.
Preferably, the attention value calculation formula is the following formula (1) or formula (2):
Uv＝a×Fv+b×Tv-c×Ev (1)
Uv＝lgFv+lgTv-lgEv (2)
the method comprises the following steps that Uv is an attention value of a vth webpage commodity to be promoted, Fv is commodity access frequency of a current vth webpage commodity to be promoted to a current user, Tv is user residence time of the current vth webpage commodity to be promoted to the current user, and Ev is commodity exposure frequency of the current vth webpage commodity to be promoted to the current user; a, b and c are pre-designated positive coefficients; 1,2,. n; and n is the number of webpage commodities to be promoted. Obviously, the attention value may also be calculated by using other formulas, as long as the formula is implemented as an increasing function of the commodity access frequency, an increasing function of the user residence time duration, and a decreasing function of the commodity exposure frequency, which are not described herein again.
S102: according to big data accumulation, determining the scores of G product classifications with the strongest relevance in a plurality of product classifications corresponding to each webpage product to be popularized, determining the scores of H webpage classification with the strongest relevance in a plurality of webpage classification corresponding to the currently visited webpage, and determining the correlation weight between each product classification in the G product classifications and each webpage classification in the H webpage classification.
Preferably, the score Rv of the G commodity classifications with the strongest relevance in the plurality of commodity classifications corresponding to each webpage commodity to be promoted1、Rv2、...、RvGCan satisfy the conditions
Score P of H webpage type classifications with strongest relevance in a plurality of webpage type classifications corresponding to currently accessed webpage1、P2、...、PHCan satisfy the conditions
Of course, the above scores may be selected not to satisfy the corresponding conditions.
In this embodiment, according to the accumulation of big data, each web page product to be promoted may be related to ten classified subject terms, the first three of the web page products may be taken out from the web page product according to the order of the degree of association (i.e., G ═ 3), each web page product may be related to ten classified subject terms, and the first three of the web page products may be taken out from the web page product according to the order of the degree of association (i.e., H ═ 3). Nine Cartesian relations are formed between the commodity classification and the webpage classification, and data of the nine Cartesian relations are obtained according to big data historical experience and are continuously corrected in the interaction behavior of the user.
For example, among ten categories corresponding to each commodity to be promoted, three categories with the strongest correlation, namely R1, R2 and R3, are selected, and through the categories of commodity characteristics and the categories of functional fields, according to the accumulation of big data, the commodity categories R1, R2 and R3 respectively have scores thereof, which are respectively: rv1、Rv2、Rv3(wherein, when a condition is defined
When the ratio is 100, Rv1+ Rv2+ Rv 3); selecting three categories P1, P2 and P3 with strongest relevance from ten categories corresponding to each webpage type, and classifying various keyword frequencies on the webpage through the attributes of the webpageThe page type classifications P1, P2 and P3 are given their scores according to the accumulation of big data, and the scores are respectively: p1、P2、P3(wherein, when a condition is defined
When it is necessary to satisfy P1+P2+P3100); there is a relative weight (the value can be limited to be between 0 and 1) for each commodity classification Ri and each webpage type classification Pi. Such as: the finance class is strongly related to the securities class, the sports class is strongly related to the Olympic class, and the related weight value is high; the culture class and the entertainment class, the life class and the education class are moderately related, and the related weight value is a medium weight value; military and living, agricultural and entertainment, etc. are weakly related, and the related weight is low. The scoring of the correlation weights between different kinds of information can be dynamically adjusted according to the result of the attention of the user.
S103: and for each webpage commodity to be promoted, determining a total commodity pushing score according to the correlation weight between each commodity classification in the G commodity classifications and each webpage type classification in the H webpage type classifications, the corresponding commodity classification score, the webpage type classification score and the attention value of the webpage commodity to be promoted.
In this embodiment, for each webpage commodity to be promoted, a plurality of total scores to be selected corresponding to each webpage commodity to be promoted are calculated according to a preset total score pushing calculation formula, and the highest total score is selected as the total score pushing the webpage commodity.
In one embodiment, the total score to be selected may be calculated in a first way as follows:
for each webpage commodity to be promoted, calculating to obtain G × H total scores to be selected of the webpage commodity to be promoted according to the formula (31) or (41):
Aij＝Uv×Rvi×Pj×Wvij(31)
Aij＝lg(Uv×Rvi×Pj×Wvij) (41)
in the formulas (31) and (41), Uv is the attention value of the vth-th webpage commodity to be promoted, RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjThe score of the jth webpage type classification in the H webpage type classifications corresponding to the currently accessed webpage is obtained; wvijThe correlation weight between the ith commodity classification in the G commodity classifications corresponding to the vth webpage commodity to be promoted and the jth webpage type classification in the H webpage type classifications; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; and n is the number of webpage commodities to be promoted.
For example: and for a certain webpage commodity to be promoted, calculating the maximum value of the comparison product according to a formula (31), and taking the maximum product value as the total pushing score of the commodity. For the classification of miss correlations, the value is set to a lowest value.
Or
In one embodiment, the total score to be selected may be calculated in the following second way:
for each webpage commodity to be promoted, calculating to obtain G × H total scores to be selected of the webpage commodity to be promoted according to the formula (32) or (42):
in the formulas (32) and (42), Uv is the attention value of the vth webpage commodity to be promoted, RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjThe score of the jth webpage type classification in the H webpage type classifications corresponding to the currently accessed webpage is obtained; wvijThe ith commodity classification in the G commodity classifications corresponding to the v webpage commodity to be promoted and the jth webpage type classification in the H webpage type classificationsCorrelation weight between classes, 0 ≦ WvijLess than or equal to 1; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; n is the number of the webpage commodities to be promoted,
when calculating the total value to be selected of the webpage commodity to be promoted, if Rvi×PjGreater than or equal to 1, then AijThe value of (a) increases exponentially; if 0 < Rvi×PjIf < 1, then AijThe value of (c) increases according to a linear law. For example: and (3) for a certain webpage commodity to be promoted, calculating the maximum value of the comparison product according to a formula (32), and taking the maximum product value as the total pushing score of the commodity. For the classification of miss correlations, the value is set to a lowest value.
Under special conditions, if the currently accessed webpage is a comprehensive portal website and the attributes of the webpage are not well classified, Rv may not be considerediAnd PjThe influence of (1) to
And Rvi×Pj×WvijAnd (4) always equals to 1, namely, the attention value of the current webpage commodity to be promoted is used as the commodity pushing total score of the webpage commodity.
S104: and sequencing the plurality of webpage commodities to be promoted from high to low according to the total score pushed by the commodities.
S105: and pushing the first N webpage commodities to be promoted in the sequencing result through N preassigned advertisement positions.
In this embodiment, the rank of the product with the highest total pushed value is higher, and for the web page with N advertisement slots, the product with N advertisement slots before the rank is pushed.
Fig. 2 is a flowchart of another advertisement push method based on big data according to an embodiment of the present invention. As shown in fig. 2, the method comprises the following steps S201-S210. Alternatively, the method comprises the following steps S201-S206, and S211-S213 (not shown):
s201: and according to the big data accumulation, determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user.
S202: according to big data accumulation, determining the scores of G product classifications with the strongest relevance in a plurality of product classifications corresponding to each webpage product to be popularized, determining the scores of H webpage classification with the strongest relevance in a plurality of webpage classification corresponding to the currently visited webpage, and determining the correlation weight between each product classification in the G product classifications and each webpage classification in the H webpage classification.
S203: and for each webpage commodity to be promoted, determining a total commodity pushing score according to the correlation weight between each commodity classification in the G commodity classifications and each webpage type classification in the H webpage type classifications, the corresponding commodity classification score, the webpage type classification score and the attention value of the webpage commodity to be promoted.
S204: and sequencing the plurality of webpage commodities to be promoted from high to low according to the total score pushed by the commodities.
S205: and pushing the first N webpage commodities to be promoted in the sequencing result through N preassigned advertisement positions.
In this embodiment, the implementation methods of steps S201 to S205 are similar to those of steps S101 to S105 in the above embodiments, and are not described herein again.
S206: and receiving selection of the webpage commodity pushed in the advertisement space.
The user clicks the content pushed in the advertisement space of the currently visited webpage, namely, the user considers that a selection instruction of webpage goods pushed in the advertisement space is received.
After step S206, the following operation one may be continuously performed, where the operation one includes the following steps S207 to S210:
s207: accessing the selected webpage commodity and timing the browsing time of the user, increasing the exposure frequency of the webpage commodity pushed in the advertisement space but not selected to the current user commodity by a first step, and increasing the commodity access frequency of the selected webpage commodity to the current user by a second step.
Wherein, for the composition of the score Uv, the value of the commodity access frequency of the current user is Fv, and the commodity access frequency is increased by a second step Sf every time the access is increased; if the commodity of the webpage commodity pushed in the advertisement space is not selected by the user, the commodity exposure frequency of the commodity to the current user is increased by the first step Se every time exposure is increased. Preferably, the commodity access frequency is updated according to the following equation (51) every time access is increased, and the commodity exposure frequency is updated according to the following equation (61) every time exposure is increased:
wherein k isE1Is a pre-specified exposure growth factor, N, greater than zeroEFor counting the number of exposures, kE2Is a preset exposure power coefficient, and 0 < kE2＜1；kF1For a pre-specified access growth factor, 0 < kE1＜kF1，NFFor the counted number of accesses, kF2For a predetermined access power coefficient, 0 < kE2＜kF2< 1, so as to ensure that the first step length of the increase of the commodity exposure frequency is far smaller than the second step length of the increase of the commodity access frequency, that is, the same commodity appears on the user page for many times, as long as the user clicks in a relatively small number of occurrences, the attention value Uv of the webpage commodity is still increased, that is, the Fv of the same webpage commodity is increased by a larger amount than that of Ev.
Preferably, the initial value of the commodity access frequency of the webpage commodity to be promoted to the current user is 0 or a first preset initial value; the initial value of the user residence time of the webpage commodity to be promoted to the current user is 0 or a second preset initial value; and the initial value of the commodity exposure frequency of the webpage commodity to be promoted to the current user is a third preset initial value. So as to adjust the values of commodity access frequency, user residence time and commodity exposure frequency in real time according to the browsing behavior of the user in the big data accumulation.
S208: and when the browsing time length is increased by a preset unit time length, increasing a third step length for the user residence time length of the current user by the selected webpage commodity.
Specifically, each time the browsing duration increases by a predetermined unit duration, the user residence duration Tv of the selected web product to the current user is updated according to the formula (71):
where Δ T is not possible because the user is infinitely resident on the current pagemaxThe preset maximum residence time is the estimated maximum residence time of the user on the visited advertisement page, such as 5 minutes; k is a radical ofTFor a predetermined attenuation factor, meaning that an increase in dwell time over time does not result in an increase in focus, e.g. kTIt may be 0.1.
S209: and receiving an access ending instruction for the selected webpage commodity.
For example, if the user clicks and closes the browsing window of the currently accessed webpage commodity through a mouse, it is considered that the access ending instruction of the selected webpage commodity is received.
S210: and ending the access to the selected webpage commodity.
In this embodiment, preferably, when step S201 is executed next time, the attention value of each webpage product to be promoted can be calculated according to the formulas (1), (51), (61), (71) as:
alternatively, after step S206, the following operation two may be continuously performed, where the operation two includes the following steps B1 to B3:
b1: and accessing the selected webpage commodity, starting timing the browsing time of the user, increasing the exposure frequency of the webpage commodity pushed in the advertisement space but not selected to the commodity of the current user by a first preset step length, and increasing the commodity access frequency of the selected webpage commodity to the current user by a second preset step length.
Wherein, for the composition of the score Uv, the value of the commodity access frequency of the current user is Fv, and the commodity access frequency is increased by a second step Sf every time the access is increased, namely the commodity access frequency is updated according to the following formula (52); if the commodity of the webpage commodity pushed in the advertisement space is not selected by the user, increasing the commodity exposure frequency of the commodity to the current user by a first step Se every time exposure is increased, namely updating the commodity exposure frequency according to the following formula (62):
Fv＝Fv+Sf (52)
Ev＝Ev+Se (62)
it should be noted here that the first step length of the increase of the commodity exposure frequency is much smaller than the second step length of the increase of the commodity access frequency, that is, the same commodity appears on the user page for many times, and as long as the user clicks in a relatively small number of occurrences, the attention value Uv of the webpage commodity is still increased, that is, the Fv of the same webpage commodity is increased by a larger amount than that of Ev.
Preferably, the initial value of the commodity access frequency of the webpage commodity to be promoted to the current user is 0 or a first preset initial value; the initial value of the user residence time of the webpage commodity to be promoted to the current user is 0 or a second preset initial value; and the initial value of the commodity exposure frequency of the webpage commodity to be promoted to the current user is a third preset initial value. So as to adjust the values of commodity access frequency, user residence time and commodity exposure frequency in real time according to the browsing behavior of the user in the big data accumulation.
B2: and receiving an access ending instruction for the selected webpage commodity.
For example, if the user clicks and closes the browsing window of the currently accessed webpage commodity through a mouse, it is considered that the access ending instruction of the selected webpage commodity is received.
B3: and ending the access to the selected webpage commodity, and updating the user residence time of the selected webpage commodity to the current user.
In this embodiment, the user residence time of the selected web product is updated according to the formula (72):
in the formula (72), T is the browsing time length recorded from the beginning of the step B1 to the step B3, T is the preset time length, T0Is a third preset step length of]Is a rounded symbol.
Preferably, after step S206, the score of the product category corresponding to the total product push score of the selected webpage product is increased, the score of the webpage type category corresponding to the total product push score of the selected webpage product is increased, and the correlation weight between the product category and the webpage type category corresponding to the total product push score of the selected webpage product is increased. For example: the scores of the three categories R1, R2 and R3 with the strongest relevance in the ten categories corresponding to the webpage commodities popularized in the current advertisement space and selected by the user are Rv respectively1、Rv2、Rv3(ii) a The scores of the three categories P1, P2 and P3 with the strongest relevance in the ten categories corresponding to the types of the web pages visited by the current user are as follows: p1、P2、P3(ii) a The related weight values of each commodity classification in R1, R2 and R3 and each webpage type classification in P1, P2 and P3 are Wv respectively11、Wv12、Wv13、Wv21、Wv22、Wv23、Wv31、Wv32、Wv33(ii) a The total goods pushing value of the selected webpage goods is Rv1、P2、Wv12I.e., the attention value Uv of the corresponding product is calculated, Rv is increased after step S2061、P2、Wv12And the value of the commodity classification and the webpage type classification corresponding to the total commodity pushing value of the webpage commodities pushed in the current advertisement space but not selectedThe score of (2), the correlation weight between the goods classification and the web page type classification decreases.
In this embodiment, the total score pushed by a commodity is implemented by the following elements: commodity access frequency, user residence time, commodity exposure frequency, commodity classification scores, webpage type classification scores, and correlation weights between each commodity classification and each webpage type classification. After the webpage commodity is pushed, relevant factors are dynamically adjusted according to the selection result of the user, so that the dynamic self-consistency of the total pushing score of the commodity under big data is achieved.
The advertisement push method based on big data provided by the embodiment of the invention is described by the specific embodiment.
Example one
Suppose there are five web page commodities to be promoted: the attention values of each commodity for the commodity access frequency, the residence time and the exposure frequency of the user are respectively as follows: the washing machine U1 is 30, the table tennis bat U2 is 30, the notebook U3 is 50, the hypoglycemic drug U4 is 60, and the down jacket U5 is 70.
With the foregoing steps S201-S206 (in which step S203 is implemented in the foregoing manner one) and steps B1-B3, the following results:
the scores of the 3 categories of the commodities with the strongest relevance corresponding to each webpage commodity to be promoted are respectively as follows:
washing machine (electric appliance R1)150, clean R1230, household R13＝20)，
Table tennis bat (sports R2150, athletics R2140, healthy R21＝30)，
Notebook (computer R3160 scientific and technical R3240, education R33＝20)，
Hypoglycemic agent (medical R4)150, healthy R4240, sanitary R43＝30)，
Down coat (clothing R5)160, fashion R5240, healthy R53＝30)。
The scores of the 3 webpage type classifications with the strongest relevance in the plurality of webpage type classifications corresponding to the currently accessed webpage are respectively as follows: health care P160, electric appliance P240, science and technology P3＝20。
The correlation weight between each commodity classification and each webpage type classification is respectively as follows:
electric appliance-health W11130, clean-healthy W12160, home-health W13140 electric appliance, electric appliance W112100, clean-electric appliance W12240 household electrical appliance W13250 electrical appliances-science and technology W11370, clean-tech W12340, household-science and technology W133＝50，
Sports-health W21170, athletics-health W22160, healthy-healthy W231Sports-electric appliance W2 ═ 1001230 sports-electric appliance W22220, health-care electric appliance W23230 sports-science W213Athletics-science W2 (50;)2340, health-science and technology W233＝50，
Computer-health W31140 science and technology-health W32150, education-health W33150, computer-electric appliance W31270 science and technology-electric appliance W32270 education-electric appliance W33240, computer-technology W31380, science and technology W323100, education-science and technology W333＝60，
Medical-health W41180, healthy-healthy W421100, sanitary-healthy W43170, medical-electric appliance W41240, health-care electric appliance W42230% sanitary electric appliance W43230, medical-science W41360, health-science and technology W41350, sanitary-science and technology W433＝30，
Clothing-health W51130, fashion-health W52140, healthy-healthy W431100, clothes-electric appliance W512Fashion-electric appliance W5 ═ 202240, health-care electric appliance W53230, clothes-Science and technology W51340, fashion-science W52350, health-science and technology W533＝50。
And then calculating the total score to be selected of each commodity to be promoted according to the formula (31):
the total score to be selected of the washing machine is that the electric appliance-health is 30 × 50 × × -2700000, the electric appliance-electric appliance is 30 2700000 3100-6000000, the electric appliance-science is 30 2700000 450 2700000 520 2700000 670-2100000, the cleaning-health is 30 2700000 730 2700000 860 2700000-2700000, the cleaning-electric appliance is 30 2700000 30 2700000 040-2700000, the cleaning-science is 30 2700000-2700000, the household-health is 30 2700000-2700000, the household-science is 30 2700000-2700000-2700000, the maximum value is 600000 obviously, and the total score pushed by the washing machine is 600000.
The total score to be selected of the ping-pong racket is 30 to 50 to 60, 30 to 150 to 240 to 1800000, 30 to 450 to 1500000, 30 to 740 to health to 740 to 960, 30 to 40 to 040 to 120, 30 to 240 to 320 to 30 to 530 to 660, 30 to 40 to 360000 to 30 to 20 to 900000, and obviously the maximum value is, so the total score of the product push of the ping-pong racket is.
The total score to be selected of the notebook computer-health is 50 × 040 ×, the computer-electric appliance is 50 × 240 ×, the computer-science and technology is 50 × 36680 ×, the science and technology-health is 50 × 740 × 860 × 950 6000000, the science and technology-electric appliance is 50 × 040 × 170 ×, the science and technology is 50 × 240 × 320 × 4100 4000000, the education-health is 50 × 520 × 660 × 50 3000000, the education-electric appliance is 50 × 20 × 40 1600000, the education-science and technology is 50 × 20 × 1200000, and the maximum value is ×, and therefore the total score of the commodity of the notebook is pushed ×.
The total score to be selected of the hypoglycemic drugs is 60 × 50 × 60 × 080-36080-14400000, 60 × 150 × 240 × 340-4800000, 60 × 450 × 520 × 660-3600000, 60 × 740 × 860 × 9100-14400000, 60 × 40 × 040 × 130-2880000, 60 × 240 × 320 × 450-2400000, 60 2400000-2400000, 60 2400000 30 2400000-2400000, and the maximum value is 2400000, so the total score to be pushed by the hypoglycemic drugs is 2400000.
The total score to be selected of the down jackets is 70-health: 60 ═ 70-electric appliance: 160 240 ═ 320 ═ 70-science: 460 640 ═ fashion-health: 70 740 940 ═ 70-040: 70-40 ═ fashion-science: 240-320 ═ health-health: 70 530 660 100 ═ health-electric appliance: 70-40 ═ 70-30 ═ health-science: 70-20 ═ 2100000, and the maximum value is obviously, so the total score of the down jackets for commodity delivery is the same.
The 5 kinds of webpage commodities to be promoted are ranked from high to low according to the total pushing score of the commodities, and the ranking sequence of the advertisements is as follows: hypoglycemic drugs, down jackets, notebooks, table tennis bats and washing machines. If the current webpage has 3 advertisement positions, 3 commodities such as hypoglycemic drugs, down jackets and notebooks are pushed.
Example two
Suppose there are five web page commodities to be promoted: the attention values of each commodity for the commodity access frequency, the residence time and the exposure frequency of the user are respectively as follows: the washing machine U1 is 30, the table tennis bat U2 is 30, the notebook U3 is 50, the hypoglycemic drug U4 is 60, and the down jacket U5 is 70.
With the aforementioned steps S201-S210 (wherein step S203 is implemented in the aforementioned manner two), the following results are obtained:
the scores of the 3 categories of the commodities with the strongest relevance corresponding to each webpage commodity to be promoted are respectively as follows:
washing machine (electric appliance R1)150, clean R1230, household R13＝20)，
Table tennis bat (sports R2150, athletics R2140, healthy R21＝10)，
Notebook (computer R3160 scientific and technical R3230, education R33＝10)，
Hypoglycemic agent (medical R4)150, healthy R4240, sanitary R43＝10)，
Down coat (clothing R5)155 fashion R5230, healthy R53＝15)。
The scores of the 3 webpage type classifications with the strongest relevance in the plurality of webpage type classifications corresponding to the currently accessed webpage are respectively as follows: health care P150, electric appliance P230, science and technology P3＝20。
The correlation weight between each commodity classification and each webpage type classification is respectively as follows:
electric appliance-health W1110.3 electrical appliance W1121 electrical appliance-science and technology W1130.7, clean-healthy W1210.6 cleaning-electric appliance W1220.4, clean-tech W1230.4, home-health W1310.4, household-electric appliance W1320.5, household-science and technology W133＝0.5，
Sports-health W211Sports-electric appliance W2 ═ 0.7120.3 sports-science and technology W2130.5, athletics-health W2210.6 athletics-electric appliance W2220.2 athletics-science and technology W2230.4, healthy-healthy W231Health-care appliance W2 ═ 1320.3, health-science W233＝0.5，
Computer-health W3110.4% W3120.7, computer-science and technology W3130.8 technical-health W3210.5 science and technology-electric appliance W3220.7 science and technology W3231, education-health W3310.5, education-electric appliance W3320.4, education-science W333＝0.6，
Medical-health W4110.8 g medical-electric appliance W4120.4, medical science and technology W4130.6, healthy-healthy W421Health-care appliance W4 ═ 1220.3, health-science and technology W4130.5, sanitary-health W4310.7, sanitary-electric appliance W4320.3, hygiene-science W433＝0.3，
Clothing-health W5110.3 clothes-electric appliance W5120.2 clothes-science and technology W5130.4, fashion-health W5210.4 fashion electric appliance W5220.4, fashion-science W5230.5, healthy-healthy W431Health-care appliance W5 ═ 1320.3, health-science W533＝0.5。
And then calculating the total value to be selected of each commodity to be promoted according to the formula (32):
total score to be selected for washing machine electric appliance-health: 30 × (50 × 50)0.3313.7 electric appliance-30 × (50 × 30)145000 electric appliance-science and technology 30 × (50 × 20)0.7238.3, clean-healthy: 30 × (30 × 50)0.62414.2, cleaning-electric appliance 30 × (30 × 30)0.4455.8, cleaning-science 30 × (30 × 20)0.4387.6, home-healthy: 30 × (20 × 50)0.4475.5 household-electric appliance 30 × (20 × 30)0.5734.8 household-science and technology 30 × (20 × 20)0.5600; it is clear that the maximum value is 45000, and therefore the total goods delivery score of the washing machine is 45000.
Total value to be selected of table tennis bat sports-health 30 × (50 × 50)0.77172.6 sports-electric appliance 30 × (50 × 30)0.3269.1 sports-science 30 × (50 × 20)0.5948.7, athletics-health: 30 × (40 × 50)0.62869.0 sports-appliance 30 × (40 × 30)0.2123.9 sports-science 30 × (40 × 20)0.4434.9, health-to-health: 30 × (10 × 50)115000, health-electric appliance 30 × (10 × 30)0.3166.0, health-science 30 × (10 × 20)0.5When 424.3, the maximum value is clearly 15000, and therefore the total value of the commercial push of the ping-pong bat is 15000.
Total score to be selected for notebook computer-health 50 × (60 × 50)0.41229.8, computer-electric appliance 50 × (60 × 30)0.79498.6, computer-science: 50 × (60 × 20)0.88719 science and technology-health 50 × (30 × 50)0.51936.5 science and technology electric appliance 50 × (30 × 30)0.75847.1 science and technology 50 × (30 × 20)130000 education-health 50 × (10 × 50)0.51118.0, education-electric appliance 50 × (10 × 30)0.4489.6 education-science 50 × (10 × 20)0.6At 1201.1, the maximum is clearly 30000, so the total item push score for the notebook is 30000.
The total score of the hypoglycemic drugs to be selected is medical treatment-health 60 × (50 × 50)0.831369.2 medical-electrical appliance 60 × (50 × 30)0.41118.4 medical-science 60 × (50 × 20)0.63785.7, health-to-health: 60 × (40 × 50)1120000 health-electric appliance 60 × (40 × 30)0.3503.4 health-science 60 × (40 × 20)0.51697.1, sanitation-health: 60 × (10 × 50)0.74649.8, sanitary-electric appliance 60 × (10 × 30)0.3332.1, sanitation-science and technology: 60 × (10 × 20)0.3294.1, apparently a maximum of 120000, so the total commercial push score for hypoglycemic agents is 120000.
Total score to be selected for Down jackets, clothing-health 70 × (55 × 50)0.3753.2 clothes-electric appliance 70 × (55 × 30)0.2308.0 clothes-science 70 × (55 × 20)0.41152.5, fashion-health 70 × (30 × 50)0.41304.8 fashion-electric appliance 70 × (30 × 30)0.41063.6 fashion-science 70 × (30 × 20)0.51714.6, healthy-healthy: 70 × (15 × 50)152500 health-electric appliance 70 × (15 × 30)0.3＝437.6，70×(15×20)0.51212.4, the maximum is clearly 52500, so the total commercial push score for the down jacket is 52500.
The 5 kinds of webpage commodities to be promoted are ranked from high to low according to the total pushing score of the commodities, and the ranking sequence of the advertisements is as follows: hypoglycemic drugs, down jackets, washing machines, notebooks and table tennis bats. If the current webpage has 3 advertisement positions, 3 commodities such as hypoglycemic drugs, down jackets and washing machines are pushed.
In the first and second embodiments, it is assumed that the user clicks on the advertisement promotion link of the hypoglycemic agent, and then pays attention to the information of the related product for a long time, the commodity access frequency value in the attention value U4 of the hypoglycemic agent is increased, the residence time value of the user is increased, and thus the U4 value is also increased. Meanwhile, the behavior of the user proves that the related weight of the medical treatment and the health can be further improved, the related weight of the classification of the hypoglycemic agent and the medical treatment can be improved, and the related weight of the classification of the webpage and the health can be improved. If the user does not click the product pushed by the corresponding advertisement, the product exposure frequency value of the user in the attention value of the product can be properly increased, and the corresponding attention value can be reduced, and meanwhile, the correlation weight, the product classification score and the webpage type classification score in the indexes with the highest correlation degrees corresponding to the product and the webpage are also reduced to different degrees.
According to the advertisement pushing method based on the big data, the webpage commodities to be promoted are scored according to the commodity access frequency of the webpage commodities to be promoted to the current user, the residence time of the user, the commodity exposure frequency, the commodity pair classification relevance, the webpage pair classification relevance and the classification relevance of the commodities to be promoted through big data accumulation, the total scores of the pushed commodities are ranked, and the pushing priorities of the commodities to be pushed are determined according to the ranked optimal results. Meanwhile, all factors influencing the total pushing score of the webpage commodity to be pushed are dynamically adjusted according to the selection of the user, so that dynamic self-consistency under big data is achieved. The method can dynamically calculate the commodities with higher interest probability of the user and push the commodities to the user, so that the aim of accurate marketing is fulfilled, and the webpage use experience of the user is improved.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (9)
1. A big data-based advertisement pushing method is characterized by comprising the following steps:
according to the big data accumulation, determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user;
according to big data accumulation, determining the scores of G product classifications with strongest relevance in a plurality of product classifications corresponding to each webpage product to be popularized, determining the scores of H webpage classification with strongest relevance in a plurality of webpage classification corresponding to the currently visited webpage, and determining the correlation weight between each product classification in the G product classifications and each webpage classification in the H webpage classification;
for each webpage commodity to be promoted, determining a total commodity pushing score according to a correlation weight value between each commodity classification in the G commodity classifications and each webpage type classification in the H webpage type classifications, and a corresponding commodity classification score, a webpage type classification score and an attention value of the webpage commodity to be promoted;
sequencing a plurality of webpage commodities to be promoted from high to low according to the total score pushed by the commodities;
the first N webpage commodities to be promoted in the sequencing result are pushed through the N preassigned advertisement positions,
the method for determining the attention value of each webpage commodity to be promoted according to the commodity access frequency, the user residence time and the commodity exposure frequency of the webpage commodity to be promoted to the current user comprises the following steps:
according to a preset attention value calculation formula, calculating the attention value of each webpage commodity to be promoted according to the commodity access frequency of the webpage commodity to be promoted to the current user, the user residence time and the commodity exposure frequency; the interest value of the commodity to be promoted in the interest value calculation formula is an increasing function of commodity access frequency, an increasing function of user residence time and a decreasing function of commodity exposure frequency.
2. The big data-based advertisement pushing method according to claim 1, wherein the attention value calculation formula is Uv-a × Fv + b × Tv-c × Ev, wherein Uv is the attention value of the v-th webpage commodity to be promoted, Fv is the commodity access frequency of the current v-th webpage commodity to be promoted to the current user, Tv is the user residence time of the current v-th webpage commodity to be promoted to the current user, Ev is the commodity exposure frequency of the current v-th webpage commodity to be promoted to the current user, a, b and c are pre-specified positive coefficients, v-1, 2.
3. The big data-based advertisement pushing method according to claim 1, wherein the attention value calculation formula is: uv ═ lgFv + lgTv-lgEv; the method comprises the following steps that Uv is an attention value of a vth webpage commodity to be promoted, Fv is commodity access frequency of a current vth webpage commodity to be promoted to a current user, Tv is user residence time of the current vth webpage commodity to be promoted to the current user, and Ev is commodity exposure frequency of the current vth webpage commodity to be promoted to the current user; 1,2,. n; and n is the number of webpage commodities to be promoted.
4. The big data-based advertisement pushing method according to claim 1, wherein for each webpage commodity to be promoted, determining a total commodity pushing score according to a correlation weight between each commodity category in the G commodity categories and each webpage type category in the H webpage type categories, and a score of the corresponding commodity category, a score of the webpage type category, and an attention value of the webpage commodity to be promoted comprises:
for each webpage commodity to be promoted, according to the formula Aij＝Uv×Rvi×Pj×WvijCalculating to obtain G × H total scores to be selected of the webpage commodities to be promoted, wherein Uv is the attention value of the vth webpage commodity to be promoted, and RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjFor the currently visited web pageThe corresponding scores of the jth webpage type classification in the H webpage type classifications; wvijThe correlation weight between the ith commodity classification in the G commodity classifications corresponding to the vth webpage commodity to be promoted and the jth webpage type classification in the H webpage type classifications; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; n is the number of webpage commodities to be promoted;
and taking the highest one of the G × H total scores to be selected of each webpage commodity to be promoted as the commodity pushing total score of the webpage commodity to be promoted.
5. The advertisement pushing method based on big data as claimed in claim 1, wherein for each to-be-promoted webpage commodity, determining a total commodity pushing score according to a correlation weight between each commodity category and each webpage type category corresponding to the to-be-promoted webpage commodity, a score of the corresponding commodity category, a score of the webpage type category, and an attention value of the to-be-promoted webpage commodity comprises:
for each webpage commodity to be promoted, according to the formula Aij＝lg(Uv×Rvi×Pj×Wvij) Calculating to obtain G × H total scores to be selected of the webpage commodities to be promoted, wherein Uv is the attention value of the vth webpage commodity to be promoted, and RviThe score, P, of the ith commodity classification in the G commodity classifications corresponding to the v-th webpage commodity to be promotedjThe score of the jth webpage type classification in the H webpage type classifications corresponding to the currently accessed webpage is obtained; wvijThe correlation weight between the ith commodity classification in the G commodity classifications corresponding to the vth webpage commodity to be promoted and the jth webpage type classification in the H webpage type classifications; 1,2, ·, G; j ═ 1,2,. H; 1,2,. n; n is the number of webpage commodities to be promoted;
and taking the highest one of the G × H total scores to be selected of each webpage commodity to be promoted as the commodity pushing total score of the webpage commodity to be promoted.
6. The big data-based advertisement pushing method according to any one of claims 1 to 5, wherein G-3 and H-3.
7. The big-data-based advertisement pushing method according to claim 4 or 5, wherein after the first N webpage commodities to be promoted in the sorting result are pushed through the N pre-specified advertisement slots, the method further comprises:
receiving a selection of the webpage commodity pushed in the advertisement space;
accessing the selected webpage commodity and starting timing the browsing time of the user, meanwhile increasing the commodity exposure frequency of the webpage commodity pushed in the advertisement space but not selected to the current user by a first preset step length, and increasing the commodity access frequency of the selected webpage commodity to the current user by a second preset step length; the first preset step length is far smaller than the second preset step length;
receiving an access ending instruction for the selected webpage commodity;
ending the access to the selected web page commodity according to the formula
Updating the user residence time of the selected webpage commodity to the current user; wherein, T is the browsing time length of this time, T is the preset time length, T0Is a third preset step length of]Is a rounded symbol.
8. The big data-based advertisement pushing method according to claim 7, wherein an initial value of a commodity access frequency of the webpage commodity to be promoted to the current user is 0 or a first preset initial value; the initial value of the user residence time of the webpage commodity to be promoted to the current user is 0 or a second preset initial value; and the initial value of the commodity exposure frequency of the webpage commodity to be promoted to the current user is a third preset initial value.
9. The big data based advertisement pushing method according to claim 7, wherein after receiving the selection of the webpage goods pushed in the advertisement slot, further comprising:
improving the commodity classification score, the webpage type classification score and the correlation weight between the commodity classification and the webpage type classification corresponding to the commodity push total score of the selected webpage commodity; and reducing the commodity classification scores, the webpage type classification scores and the correlation weights between the commodity classifications and the webpage type classifications corresponding to the commodity pushing total scores of the webpage commodities pushed but not selected in the advertisement space.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710120950.1A CN107103487B (en) | 2017-03-02 | 2017-03-02 | Advertisement pushing method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710120950.1A CN107103487B (en) | 2017-03-02 | 2017-03-02 | Advertisement pushing method based on big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107103487A CN107103487A (en) | 2017-08-29 |
| CN107103487B true CN107103487B (en) | 2020-07-14 |
Family
ID=59675538
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710120950.1A Active CN107103487B (en) | 2017-03-02 | 2017-03-02 | Advertisement pushing method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107103487B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798560A (en) * | 2017-10-23 | 2018-03-13 | 武汉科技大学 | A kind of retail shop's individual character advertisement intelligent method for pushing and system |
| CN107895299B (en) * | 2017-11-08 | 2020-12-01 | 浙江执御信息技术有限公司 | Commodity exposure sequencing method and device |
| CN108400929B (en) * | 2018-02-08 | 2021-06-08 | 网易乐得科技有限公司 | Data processing method, device, computing equipment and medium |
| CN108600325B (en) * | 2018-03-27 | 2021-08-10 | 努比亚技术有限公司 | Push content determining method, server and computer readable storage medium |
| CN108830492B (en) * | 2018-06-22 | 2022-02-08 | 成都博宇科技有限公司 | Method for determining spot-check merchants based on big data |
| CN109343765B (en) * | 2018-08-16 | 2021-03-23 | 咪咕数字传媒有限公司 | Page turning method of electronic book, reading equipment and storage medium |
| CN110413870B (en) * | 2018-12-18 | 2021-12-31 | 北京沃东天骏信息技术有限公司 | Commodity recommendation method and device and server |
| CN111582956A (en) * | 2020-06-29 | 2020-08-25 | 成都新潮传媒集团有限公司 | Advertisement pushing method and device and computer equipment |
| CN113743972B (en) * | 2020-08-17 | 2024-06-18 | 北京沃东天骏信息技术有限公司 | Method and device for generating article information |
| CN113407858B (en) * | 2021-08-20 | 2021-12-21 | 明品云(北京)数据科技有限公司 | Information push method, system, medium and electronic device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103995839A (en) * | 2014-04-30 | 2014-08-20 | 兴天通讯技术（天津）有限公司 | Commodity recommendation optimizing method and system based on collaborative filtering |
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
-
2017
- 2017-03-02 CN CN201710120950.1A patent/CN107103487B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103995839A (en) * | 2014-04-30 | 2014-08-20 | 兴天通讯技术（天津）有限公司 | Commodity recommendation optimizing method and system based on collaborative filtering |
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107103487A (en) | 2017-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107103487B (en) | Advertisement pushing method based on big data | |
| Webster et al. | The Internet audience: Web use as mass behavior | |
| US7882175B1 (en) | Selecting an advertising message for presentation on a page of a publisher web site based upon both user history and page context | |
| US8364627B2 (en) | Method and system for generating a linear machine learning model for predicting online user input actions | |
| CA2700030C (en) | Touchpoint customization system | |
| US8725752B2 (en) | Method, apparatus and computer readable medium for indexing advertisements to combine relevance with consumer click feedback | |
| US8321278B2 (en) | Targeted advertisements based on user profiles and page profile | |
| Dembczynski et al. | Predicting ads click-through rate with decision rules | |
| CN102855256B (en) | For determining the method, apparatus and equipment of Website Evaluation information | |
| US20120078710A1 (en) | Dynamic search engine results employing user behavior | |
| EP2159748A1 (en) | System and method for providing topic-guided broadening of advertising targets in social indexing | |
| US20130290303A1 (en) | Categorizing Documents | |
| CN105917364B (en) | Ranking discussion topics in question-and-answer forums | |
| CN109993583B (en) | Information pushing method and device, storage medium and electronic device | |
| CN108509466A (en) | A kind of information recommendation method and device | |
| EP2156294A1 (en) | Probabilistic recommendation system | |
| WO2010085773A1 (en) | Hybrid contextual advertising and related content analysis and display techniques | |
| JP2011519080A (en) | Method and apparatus for selecting related content for display in relation to media | |
| CN106682926A (en) | Method and apparatus for pushing search advertisements | |
| CN110275952A (en) | News recommended method, device and medium based on user's short-term interest | |
| US20130006760A1 (en) | Systems and methods for presenting comparative advertising | |
| CN113239182A (en) | Article recommendation method and device, computer equipment and storage medium | |
| CN109446402B (en) | Searching method and device | |
| CN113961823B (en) | News recommendation method, system, storage medium and equipment | |
| Xie et al. | A probabilistic recommendation method inspired by latent Dirichlet allocation model |
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 |