TWI499990B - Method, device and system for improving the transmission speed of website data - Google Patents

Description

Method, device and system for improving website data transmission speed

The present invention relates to the field of network technologies, and in particular, to a method, device and system for improving the speed of data transmission of a website.

At present, with the increasing variety of Internet services, the amount of data transmitted by the web server to the client is increasing, for example, various graphic materials, voice data, video materials, and the like. When such a large amount of website data is transmitted to the client at the same time, it will lead to a sharp drop in the speed of network data transmission, and even cause the embarrassment of the entire website. In the case of online advertising, online advertising can quickly convey business information to the user community and stimulate the user's desire to purchase. Therefore, when a user browses a website, the website server usually transmits some network advertisements to the user terminal. According to the information, if a large number of users browse the website at the same time, the website server will transmit a large amount of advertisement data to the client at the same time, which causes the network data transmission speed to slow down, and even causes the website server to be paralyzed. In order to reduce the impact of the transmission of online advertising materials on the transmission speed of the network, in the prior art, the effect of increasing the speed of data transmission on the website is often achieved by reducing the amount of advertising data transmitted to the client. However, blindly reducing the amount of advertising data. Undoubtedly will reduce the effectiveness of advertising. How to improve the transmission of website advertisements on the basis of ensuring the effectiveness of advertisements has become an important issue to be solved urgently.

The embodiment of the invention provides a method, a device and a system for improving the data transmission speed of the website, which are used to reduce the amount of data transmitted during the advertisement delivery on the basis of ensuring the effect of the advertisement delivery.

The specific technical solution provided by the embodiment of the present invention is as follows: a method for improving the data transmission speed of a website, comprising: obtaining a corresponding feature attribute set according to an operation behavior when a user browses a website, and then collecting the feature attribute in the preset rule base according to the feature attribute set; Obtaining at least one rule that matches the feature attribute set; filtering at least one advertisement corresponding to the scene restricted by the rule according to the obtained at least one rule, and delivering the at least one advertisement to the user; monitoring the user for the at least one The operational behavior generated by the delivery of an ad, and the relevant parameters collected are converted into corresponding rules to update the rule base.

An apparatus for improving data transmission speed of a website, comprising: an obtaining unit, configured to obtain a corresponding feature attribute set according to an operation behavior when a user browses a website, and obtain the feature in the preset rule base according to the feature attribute set; At least one rule matched by the set of attributes; the first processing unit is configured to filter at least one advertisement corresponding to the scene restricted by the rule according to the obtained at least one rule, and deliver the at least one advertisement to the user; a unit, configured to monitor an operation behavior generated by the user for the delivery of the at least one advertisement, and convert the collected related parameters into corresponding rules to update the rule base.

A system for improving the speed of data transmission of a website, comprising: a rule base for storing various rules for searching for advertisements; and an advertisement management device for obtaining a corresponding feature attribute set according to an operation behavior when the user browses the website, and Acquiring at least one rule matching the feature attribute set in the preset rule base according to the feature attribute set, and filtering at least one advertisement corresponding to the scene restricted by the rule according to the obtained at least one rule, and The user delivers the at least one advertisement, and monitors the operational behavior generated by the user for the delivery of the at least one advertisement, and converts the collected related parameters into corresponding rules to update the rule base.

In the embodiment of the present invention, in order to accumulate a good delivery experience, the concept of a rule base is introduced, which is classified according to various factors of the delivery association, and the effect of each category is provided. The better part is statistically summarized, and some of the better delivery matching rules in each type of delivery are summarized. The establishment and evolution of the rule base are directly based on the effect of the advertisement delivery, and the effect of the advertisement delivery changes. Instantly responding to the various rules that are stored to guide the selection of advertisements, making the selection of advertisements completely depend on the effect of their advertisements, and also making the update of the rule base evolve based on the effects of advertisements, making the most of the rules. Jiahua can be automated and instant, with the advantages of low cost, short cycle and fast optimization. In this way, there is no need to blindly reduce the amount of advertising, but the targeted advertising is targeted according to the actual needs of the user, and the amount of unnecessary advertisements is reduced, thereby reducing the effectiveness of the advertising. Information transmitted when a website ad is served The quantity increases the data transmission speed of the system, thereby improving the service quality of the website.

In this embodiment, the rule base based on the advertisement effect is used to support the selection of the advertisement delivery strategy to improve the transmission speed of the website data. Specifically, the device for managing advertisement delivery obtains a corresponding feature attribute set according to an operation behavior when the user browses the website (for example, a scene in which the user browses the webpage at the time), including a browsing time period, a browsed webpage ID, and an advertisement location ID, and a user identifier ID or the like, and acquiring at least one rule matching the feature attribute set in the preset rule base according to the feature attribute set, and filtering corresponding to the scene restricted by the rule according to the obtained at least one rule At least one advertisement, and the at least one advertisement is served to the user, and the operation behavior generated by the user for the delivery of the at least one advertisement is monitored, and the collected related parameters are converted into corresponding rules to update the rule base. The feature attribute set is used to describe the particularity of the user's browsing time, the characteristics of the browsed webpage and the advertisement, the long-term interest preference of the user, and the recent operational behavior intention preference of the user when browsing the website. In this way, there is no need to blindly reduce the amount of advertising, but the targeted advertising is targeted according to the actual needs of the user, and the amount of unnecessary advertisements is reduced, thereby reducing the effectiveness of the advertising. The amount of data transmitted by the website advertisements at the time of delivery increases the speed of data transmission of the system, thereby improving the service quality of the website.

The so-called advertising effect, that is, after the ad is delivered, it is used to measure A metric that is popular with users, including a variety of preset parameters, such as user click-through rate, pageviews after reaching the target page, registration amount, collection amount, purchase amount, and many other indicators.

Rule base: refers to the results from past advertising, based on the classification of many factors related to the delivery, statistical summary of the delivery of each category, and summed up some of the better delivery of each type of advertising A total set of matching rules for which the experience of genetic evolution needs to be accumulated on an ongoing basis, and accumulated experience is used to guide future advertising.

The preferred embodiments of the present application are described in detail below with reference to the accompanying drawings.

Referring to FIG. 1 , in the embodiment of the present application, a system for managing advertisement delivery to improve website data transmission speed includes a rule base 10 and an advertisement delivery management device 11 .

The rule base 10, which is used to store various rules for searching for advertisements, is a cumulative experience of all advertising delivery policy implementations, and is always undergoing an instant evolution update. The accumulation of various rules in the rule base 10 allows the implementation of an effective advertising strategy to be preserved, thereby providing valuable experience for subsequent operations. In this embodiment, when formulating an advertisement delivery strategy for placing an advertisement, all types of factors affecting the effect of the advertisement delivery are comprehensively considered, and the selection of the advertisement delivery strategy is performed at one time, thereby ensuring the global optimization of the advertisement delivery strategy. For example, when selecting an advertisement delivery strategy for an advertisement, various parameters in the advertisement delivery strategy are set according to characteristic information such as the current advertisement location, the delivery scenario, the browsing user's interest, and the recent browsing behavior, for example, Delivery time, number of deliveries, and more.

The advertisement distribution management device 11 is configured to obtain a corresponding feature attribute set according to an operation behavior when the user browses the website, and obtain at least one rule matching the feature attribute set in the preset rule base according to the feature attribute set, and then And filtering at least one advertisement corresponding to the scene restricted by the rule according to the obtained at least one rule, and delivering the at least one advertisement to the user, and monitoring an operation behavior generated by the user for the delivery of the at least one advertisement, and collecting The relevant parameters are converted to the corresponding rules to update the rule base.

In the embodiment of the present application, when the advertisement delivery policy is selected, the advertisement delivery strategy adopted by the same or similar delivery instance may be searched as a reference material, and the delivery rule corresponding to the advertisement effect of the delivery instance is determined according to the delivery performance score. Sort from the largest to the smallest, find out the optimal advertising effect and corresponding advertising feature parameters, and make appropriate combinations of these advertising feature parameters to mutate or expand the variability, and then based on the mutated advertising The characteristic parameters select qualified advertisements that meet the conditions, and perform a campaign campaign for each candidate advertisement according to the comprehensive score of the delivery effect, filter out the finally delivered advertisements, and then immediately track the delivered advertisements to monitor the delivery effect. Finally, the ad delivery strategy currently selected according to the delivery effect is adjusted and updated, the good delivery mode is accumulated, and the poor delivery mode is discarded, so that the advertisement delivery strategy is optimized. In this way, the amount of data transmitted by the network advertisement on the network is reduced, and the good advertisement delivery effect can be received.

Referring to FIG. 2, in the embodiment of the present application, an advertisement delivery management device 11 includes an obtaining unit 110, a first processing unit 111, and a second processing unit 112, where

The obtaining unit 110 is configured to obtain a corresponding feature attribute set according to an operation behavior when the user browses the website, and then acquire at least one rule matching the feature attribute set in the preset rule base according to the feature attribute set.

The first processing unit 111 is configured to filter at least one advertisement corresponding to the rule-constrained scene according to the obtained at least one rule, and deliver the at least one advertisement to the user;

The second processing unit 112 is configured to monitor an operation behavior generated by the user for the delivery of the at least one advertisement, and convert the collected related parameters into corresponding rules to update the rule base.

In the embodiment of the present application, in the rule base 10, a rule is composed of the following data vectors, including:

A, ad slot feature vector (referred to as F _a), the component comprising: advertising a position corresponding to the channel site (referred to as ), advertising location category (denoted as ), the category of the page where the ad location is located ( ), the keyword of the page where the ad is located ( ). The relationship of the above parameters can be expressed as: .

B. The advertisement location is given a scene feature vector (denoted as F _b ), and the components included are: the delivery period (denoted as ), date type (marked as ), season (recorded as ), current events mark (marked as ), where the current events are used to mark whether there is a major event recently. The types of so-called major events include but are not limited to: earthquake, politics, economics, college entrance examination, medical treatment, and so on. The relationship of the above parameters can be expressed as: .

In the embodiment of the present invention, the new vector F _ab =( F _a , F _b ) generated by the vector F _a connection vector F _{b is} called an advertisement location vector, and the vector describes that the advertisement is not dependent on the overall delivery effect of the user. factor.

C, user natural attributes and historical long-term interest behavior feature vectors (denoted as F _c ), including the components: user gender (marked as ), user age (recorded as ), user interest (marked as , that is, the user's daily online rules, holiday, time), user shopping interest (recorded as , that is, the category of goods that users browse and purchase on a daily basis), keywords that users like (marked as ), user brand tendency (denoted as ), user consumption level (marked as , that is, the price segment of the goods that the user browses and purchases), the tendency of the user's business (denoted as ), user area (recorded as And user credit (denoted as ). The relationship of the above parameters can be expressed as:

D, the user's recent instant browsing and shopping feature vector (denoted as F _d ), including the short-term and current click advertising categories (denoted as ), short-term and current browsing categories (denoted as ), short-term and current categories of goods purchased (marked as ), short-term and current click ad location categories (denoted as ), short-term and current browsing of web categories (denoted as ). The relationship of the above parameters can be expressed as: .

Embodiments of the present invention, the connection of a new vector the vector F _c F _d F _cd vector generated = (F _c, F _d) a feature vector called a user, on behalf of the user himself short and long term characteristic properties, also called user feature attribute vector.

E, the advertisement location advertisement delivery strategy feature vector (denoted as F _e ), and the included component has an advertisement delivery strategy (denoted as ) and the corresponding configuration parameters (marked as ). The advertisement delivery strategy is a delivery method used when the advertisement is displayed, for example, using a keyword-content matching algorithm, using a user-behavior matching algorithm to deliver or according to an advertisement effect, and corresponding to an advertisement delivery strategy. Configuration parameters, which can include user IDs, ad keywords, and more. The relationship of the above parameters can be expressed as:

F, the advertising feature vector (denoted as F _f ), the components included are: the type of advertising product (marked as ), advertising categories (remarked as ), the form of advertising display (marked as , that is, graphic, text link, or flash), advertising content custom parameters (marked as , the keyword used for clicks, etc.), the auction’s bidding keyword (marked as ), the bid price of the ad (marked as ), the creditworthiness of the advertiser (marked as ), the brand of advertising goods (marked as ), the price segment of the advertising product (marked as ), advertising business type (marked as ), advertising business area (recorded as ). The relationship of the above parameters can be expressed as:

In the embodiment of the present application, the vectors F _a , F _b , F _c , F _d , F _e , F _{f are} connected to generate a new vector F = ( F _a , F _b , F _c , F _d , F _e , F _f ) This vector is a detailed description of the rule base used to develop the ad serving strategy.

G, the advertising effect is normalized to the indicator vector (denoted as F _g ), and the components included are: click rate (denoted as ), click on the income (recorded as ), introduce traffic (denoted as ), the number of collections ), the transaction amount (denoted as ), commission amount (recorded as ), the transaction rate (denoted as And registration rate (denoted as )

Through the vector F _g , we can calculate the score S for describing the effect of the advertisement, and the calculation formula of S is as follows: among them, , w _i is called the weight coefficient; Is a normalized function, will Converted to a value between 0 and 100.

Therefore, the range of S is 0-100, and the weight coefficient w _i is preset by the administrator based on the experience value, for example, the click rate is confirmed. Is the most important factor for measuring the performance of an ad, so you can preset w ₁ =1, then For another example, to confirm that all components of F _g are equally important, then w _i = 1/8 = 0.125 can be preset. In short, w _{i is} closer to the value 1, it means The weight of the corresponding component in measuring the performance of the ad is greater.

In the embodiment of the present application, the vectors F _a , F _b , F _c , F _d , F _e , F _f , F _{g are} connected into a new vector F _stat =( F _a , F _b , F _c , F _d , F _e , F _f , F _g ), the vector F _{stat is} called the advertising effect statistics indicator vector.

Based on the above parameter settings, a specific application scenario is taken as an example for detailed description. Assume that there are three ads initially selected for delivery, namely Ad A, Ad B, and Ad C. After the three advertisements are served for a certain period of time, when a user logs in to the website, the system needs to determine which advertisement to select to serve to the login user according to the effect of the three advertisements.

In this embodiment, the preset rules and user access conditions in the rule base are assumed. The scenery is as follows:

Three advertisements A, B, C:

Advertisement A: The advertising product MP3, the price of the advertising product is <1000 yuan, the credit of the owner is 200 points, the advertisement is displayed by the picture, the keyword is selected to match exactly, and the bid price is 0.3 yuan.

Advertising B: advertising product touch mobile phone, advertising product price > 2000 yuan, store owner credit 500 points, advertising using flash display, select keywords fuzzy match delivery, bid price 0.8 yuan.

Advertising C: Advertising product dolls, the price of advertising products is <100 yuan, the credit of the owner is 30 points, the advertisements are displayed by pictures, the keywords are selected to be fuzzy matching, and the bidding price is 1 yuan.

The above advertisements are distributed by the management personnel on the network side, and are pre-stored in the database and obtained by the advertisement search engine. For the above three advertisements, the following six rules are preset in the rule base:

1, R1 = (male users, interested in digital products, income in the middle, recently purchased touch phones, often access to news ads, click on the ads are MP3, advertising product prices <2000 yuan, advertising time for the weekend, delivery The advertiser's credit for the advertisement is greater than 20 points, the advertisement is displayed in flash, and the advertisement is delivered in the exact way of the keyword, 0.2 yuan <click auction price average price <0.4 yuan).

2, R2 = (male users, interested in sports equipment, income unknown, recently purchased roller skates, often visit blog advertising positions, click ads are touch phones, advertising prices > 2000 yuan, advertising time is week At the end of the morning, the advertiser’s credit for placing the advertisement is greater than 300 points, the advertisement is displayed in flash, and the advertisement is placed in a keyword fuzzy manner, 0.3 yuan <click auction price average price <1 yuan)

3, R3 = (male users, interested in sports equipment, no income (student), recently purchased perfume, often visit anime advertising position, click on the ad is a doll, advertising price <100 yuan, advertising time is a working day In the evening, the advertiser’s credit for placing the advertisement is greater than 20 points, the advertisement is displayed in the image, and the advertisement is placed in a keyword fuzzy manner, 0.3 yuan <click auction price average price <1.3 yuan).

4, R4 = (female users, interested in sports equipment, high income, recently purchased perfume, often visit news advertising positions, click on the ads are touch phones, advertising products prices > 5000 yuan, advertising time is the working day morning The advertiser's credit for placing the advertisement is greater than 500 points, the advertisement is displayed by the image, and the advertisement is delivered in the exact way of the keyword, 0.3 yuan <click auction price average price <1.3 yuan)

5, R5 = (female users, interested in dolls, income, recently purchased MP3, often visit blog ads, click on the ad is a doll, advertising product price <100 yuan, the delivery period is weekend night, the advertisement is placed The advertiser credit is greater than 30 points, the advertisement is displayed in the image, and the advertisement is placed in the exact way of the keyword, 0.5 yuan <click auction price average price <0.8 yuan).

6, R6 = (female users, interested in decorations, income in the middle, recently bought MP3, often visit anime ads, click on the ad is touch phone, advertising product price > 2000 yuan, the launch time is the weekend morning The advertiser's credit for placing the advertisement is greater than 300 points, the advertisement is displayed by the image, and the advertisement is placed in a keyword fuzzy manner, 0.5 yuan <click auction price average price <0.8 yuan)

Based on the above rules, assume that the user access scenario is as follows:

Scenario 1: (User U1, frequently accessing news advertising locations on weekend mornings)

Scenario 2: (User U2, often visits blog-type advertising locations on weekday nights)

Scenario 3: (User U3, frequently accessing news advertising locations on weekday mornings)

The advertisement distribution management device 11 collects the user's access information according to the above three scenarios, and stores the access information in the website log, and after analyzing the website log, extracts the feature attribute set of each user.

Then, you can get the set of feature attributes of these three users, which are:

The characteristic attribute of user U ₁ is (male, interested in digital products, income in the middle, recently purchased touch phones)

The characteristic attribute of user U ₂ is (female, interested in doll equipment, income, recently purchased MP3)

The characteristic attributes of user U ₃ are (female, interested in sports equipment, high income, recently purchased touch phones).

Then, referring to FIG. 3, in the embodiment of the present application, the detailed process of the advertisement delivery management device 11 for managing the advertisement delivery based on the advertisement delivery effect is as follows:

Step 300: After determining that a user logs in to the website system, obtain a corresponding feature attribute set according to the operation behavior of the user when browsing the website, and select a matching rule according to the feature attribute set in a preset rule base, where the rule is used for Select an alternate ad that matches the feature attributes of the user.

For example, access to user U ₁ (male, interested in digital products, revenue in the middle, recent purchases of touch phones, visits during weekends, frequent visits to news advertising positions), through the function H _similarity ( U ₁ , F _i) may be calculated for all rules 10 and U values similarity rule base _1, and then sorted by inverting similarity values, according to a set threshold, taking the ranking, which is found in the rules in the rule base TopX rule of the user The rule that U1 has the same or similar feature attributes.

Where x, y F, F = ( F _a , F _b , F _c , F _d , F _e , F _f ), i takes the value [a, f], F ₀ ~ F _f is the default used in the rule base to describe the various types The set of advertising attributes, F ₀ ~ F _{f is} used to form F _i , and j is the component contained in F _i .

Of course, the above F = ( F _a , F _b , F _c , F _d , F _e , F _f ) is only an example. In practical applications, if more based on the actual application environment, more defined vector sets are added, such as F. =( F ₁ , F ₂ ,..., F _n ), F _a , F _b , F _c , F _d , F _e , F _f are six of them, then the above formula The same applies, where x, y F, F = ( F ₁ , F ₂ , ..., F _n ), i takes the value [1, n], and F ₀ ~ F _n are presets in the rule base to describe various types of advertising attributes. The set of F ₀ ~ F _{n is} used to form F _i , and j is the component contained in F _i .

After using the retrieval function H _similarity , the rule R1 can be selected for the user U1: (male user, interested in digital products, income in the middle, recently purchased touch mobile phone, frequent access to news advertising positions, clicked ads are MP3, advertising The product price is <2000 yuan. The delivery period is weekend. The advertiser's credit for placing the advertisement is greater than 10 points. The advertisement is displayed in flash, and the advertisement is delivered in the exact way of the keyword. 0.2 yuan < click auction price average price <0.4 yuan).

In the actual application, the rule to be selected may be one or multiple. In the embodiment of the present application, it is assumed that the selected rule matching the user feature set of the login user is R4, R5, and R6.

Step 310: Filter out corresponding candidate advertisements according to the selected rule.

For example, suppose the selected rule matching the set of user features is R4, R5, R6, and then the user ID and the keyword extracted from the selected rule are passed as parameters to the advertisement search engine for searching by the advertisement. The engine searches for the corresponding alternate ad based on the parameters obtained. In this embodiment, it is assumed that the advertisement search engine selects the corresponding candidate advertisements as advertisement A, advertisement B, and advertisement C according to the selected rules R4, R5, and R6.

Step 320: Perform a probability campaign for the obtained alternative advertisement.

In the embodiment of the present application, the candidate advertisement is subjected to a probabilistic campaign in the following manner:

The alternative advertisements filtered according to rules R4, R5 and R6 are expressed as: Where i is the corresponding rule subscript and j is the number of alternative advertisements obtained. In this embodiment, the value of i is 4, 5, 6. Then, all the filtered alternative ads are expanded as follows:

The probability campaign step is as follows: according to the calculated probability of the campaign score H _result , the selected rule Ri is reverse sorted, and the function used is H _result ( x , y )= e ^βS × H _similarity ( x , y ), wherein β is the preset effect expansion factor, initially set to 1, the manager can optimize according to the test effect selected by the β parameter, S is the effect score of the rule corresponding to y, x , y F _abcd , F _abcd =( F _a , F _b , F _c , F _d ), x represents a connection vector of the advertisement position vector F _ab corresponding to the user-specific access and the user feature vector F _cd , belonging to F _abcd , y indicating The connection vector of the advertisement position component F _ab and the user feature component F _cd in the rule R is also attributed to F _abcd .

Select the TopX (the result of the top X) of the sorted Ri, and then determine the corresponding candidate advertisement for the selected TopX. Here, assuming X=2, then the final rule is R4, R5, and corresponding The alternative ads are Ad A and Ad B, expressed as , , abbreviate this ad collection as Ad.

Finally, the random sampling is performed again for the set Ad, and the number of samples is Y (according to the parameter setting of the system, assuming Y=1), then the final probability election result may be advertisement A or advertisement B.

Step 330: Perform delivery presentation on the finally selected advertisement.

Step 340: Monitor the user's action line for the final delivered ad. The rule base 10 is updated according to the collected advertisement effect data.

In the above step 340, after the final selection of the advertisement is displayed, the log generated by the delivery is collected in step 350. The main contents of the log record include but are not limited to: user id, access time, clicked ad position, browsed ad position, and favorite products.

After a certain period of time from the delivery time, the effect of the advertisement on the advertisement is calculated, specifically calculating the performance data of the advertisement (including the effect score S and the support degree N), and then based on the calculated advertisement delivery effect data, the rule base The rule saved in 10 is updated. In the embodiment of the present application, the update of the rule base 10 includes two operations: 1. The corresponding new rule is extracted according to the advertisement delivery effect data and added to the rule base 10, 2. According to the advertisement The delivery performance data optimizes the existing rules in the rule base 10.

The so-called extraction means that a large number of occurrences (ie, the probability is greater than a certain threshold) of the advertising effect statistical indicator vector F _{stat is} converted into a rule.

For example, the user U accesses a specific webpage W, an advertisement location P on the webpage, and an advertisement A on the advertisement location P. After the user sees the advertisement A, the user clicks on the link of the advertisement A. , enter the product details page P of the advertisement A for promotion, and then purchase the product I on the detail page P, and collect the product J. The user's operation behavior of this series will be recorded by the system as: (U, T, W, P, A, I, J), see the above-mentioned set C and set D; Then, the recorded series of operation behaviors of the user U are collated and analyzed, and the corresponding content is saved as the feature attribute set of the user. Including: T is converted into a corresponding time period Ti, the season Ts of the delivery, whether there is a major holiday Tf, and the like; Then, W and P are converted into the advertisement location feature data set required by the rule base 10 through the advertisement location data in the customer relationship management (CRM) system and the advertisement location text data in the existing advertisement search engine. See the aforementioned set A; Finally, through the advertising materials in the advertising CRM system, and through the advertiser's promotion product system, the detailed attributes of A and I are obtained, and the two are combined and converted into the delivered advertising feature data (see the above-mentioned collection F).

Thus, the user's series of operational behaviors (U, T, W, P, A, I, J) will be converted to the aforementioned advertising effect statistical indicator vector F _stat .

Then according to the formula Calculated advertisement effect statistics index vector F _stat effect score S _new and F _stat of support N _new, if s _new> set threshold value, and N _new> the set threshold value, if the F _stat does not exist in the rules database 10, Then, F _{stat is selected} as the extracted new rule into the rule base 10, thus completing the extraction of the new rule.

If F _stat already exists in the rule base 10, the effect score of the original stored vector F _stat in the rule base 10 is recorded as S _old , the support degree record is N _old , and then the combined effect score is calculated according to the following formula: S _Combination = α × S _old +(1- α ) S _new N _combination = β × N _old +(1- β ) N _new

According to the results, if S _engagement> set threshold, and _the N> set threshold S _Old update will rule base 10 of the original stored rules F _stat is S _engagement, is updated N _old is _the N; if S _bonded < When the threshold is set, or N _is <set threshold, the corresponding rule F _stat is deleted from the rule base 10, thus completing the optimization of the existing rule.

The calculation function of the above support degree N is Support ( x ): , wherein, in a certain period of time, the recorded F _stat vector set is recorded as SetF , x F _stat .

On the other hand, in the above embodiment, in order to perform the genetic variation on the selected rule after the execution of step 300, it is preferable to add a new rule in the rule base 10. The selected rules can be genetically mutated, or the selected rules can be randomly sampled, and the extracted rules are genetically mutated.

In the embodiment of the present application, the genetic variation method that can be used includes, but is not limited to, cross-mutation using the genetic algorithm to select the rule selected in step 300, which is specifically: the rule for performing genetic variation is R4=( F _a , F _b , F _c , F _d , F _e , F _f , F _g ), and R5=( F _a , F _b , F _c , F _d , F _e , F _f , F _g ) Then, first, the rule R4 R5 is encoded and can be encoded naturally.

Next, select the mutation points of rules R4 and R5. In order to reduce the variation and produce a large number of useless offspring, the variation point can be selected as the position between F _d and F _e , as shown by the double vertical line below: ( F _a , F _b , F _c , F _d ∥ F _e , F _f , F _g )

So, it may be _{R4 = (F a, F b} , F c, F d, F e, F f, F g) according to the position of point mutation, split into _{:( F a, F b, F} c, F _d ) and ( F _e , F _f , F _g )

Then, the split vectors are cross-connected: ( F _a , F _b , F _c , F _d ) and ( F _e , F _f , F _g ) ^' connected ( F _a , F _b , F _c , F _d , ( F _e , F _f , F _g ) ^' ) ( F _a , F _b , F _c , F _d ) ^' and ( F _e , F _f , F _g ) ^{' are} connected (( F _a , F _b , F _c , F _d ) ^' , F _e , F _f , F _g )

In this way, new rules ( F _a , F _b , F _c , F _d , ( F _e , F _f , F _g ) ^' ) and (( F _a , F _b , F _c , F _d ) after genetic variation are obtained. ^' , F _e , F _f , F _g ).

In the above embodiment, the genetic variation of the existing rules can give the "variation" of the appropriate probability of the advertisement placement strategy while selecting the top optimal rule based on the historical effect, and these variations ensure the "evolution" of the rule base 10. You can discover and mine new rules, which will help you expand your advertising model.

In summary, in the embodiment of the present invention, in order to improve the experience of good delivery Line accumulation introduces the concept of rule base 10, which is based on the many effects brought by past advertisements, classified according to many factors of delivery association, and statistically summarizes the better parts of each category. Some of the best match-matching rules in each type of delivery, and continuous genetic evolution on the line, accumulate to guide the update of the rule base 10 in the future. In this way, the advertisement placement based on the rule base 10 is made simple and easy, and the global optimization of the advertisement delivery can be better achieved. On the other hand, in addition to online guidance, the rule base 10 is also a summary of experience, offline can guide the development and innovation of the business.

The establishment and evolution of the rule base 10 are directly based on the effect of the advertisement delivery, and the effect of the advertisement delivery changes, and will immediately reflect through the rule base 10 on various rules for saving advertisement selection, so that the advertisement The choice is completely dependent on its delivery effect, so that it is formed: advertising delivery - delivery effect tracking - rule update - advertising re-delivery, such a large delivery loop, so that the purpose and means are unified. In short, the update evolution of the rule base 10 is realized based on the effect of the advertisement in real time, which can automate and instantiate the optimization of various rules, and has the advantages of low cost, short cycle and fast optimization. . In this way, there is no need to blindly reduce the amount of advertising, but the targeted advertising is targeted according to the actual needs of the user, and the amount of unnecessary advertisements is reduced, thereby reducing the effectiveness of the advertising. The amount of data transmitted by the website advertisements at the time of delivery increases the speed of data transmission of the system, thereby improving the service quality of the website.

Obviously, those skilled in the art can advance the embodiments of the present invention. Various changes and modifications may be made without departing from the spirit and scope of the invention. Accordingly, the present invention is intended to cover such modifications and variations as the modifications and variations of the embodiments of the present invention are intended to be included within the scope of the invention.

10‧‧‧ rule base

11‧‧‧Advertising Management Device

12‧‧‧Ad search engine

110‧‧‧Acquisition unit

111‧‧‧First Processing Unit

112‧‧‧Second processing unit

FIG. 1 is a schematic structural diagram of an advertisement delivery management system according to an embodiment of the present application; FIG. 2 is a functional structural diagram of an advertisement delivery management apparatus according to an embodiment of the present application; FIG. 3 is a flowchart of managing an advertisement delivery based on an advertisement delivery effect according to an embodiment of the present application; Figure.

Claims (10)

A method for improving the speed of data transmission of a website, comprising: obtaining a corresponding feature attribute set according to an operation behavior when a user browses a website, and acquiring, according to the feature attribute set, a set of the feature attribute in a preset rule base Matching at least one rule; filtering at least one advertisement corresponding to the scene restricted by the rule according to the obtained at least one rule, and delivering the at least one advertisement to the user; monitoring an operation generated by the user for the delivery of the at least one advertisement Behavior, and the relevant parameters collected are converted into corresponding rules to update the rule base. The method of claim 1, wherein, when the at least one rule matching the feature attribute set is obtained in the preset rule base according to the feature attribute set, the method includes: calculating each rule in the rule base The similarity with the feature attribute set; sorting the rules according to the similarity degree from the largest to the smallest; starting from the rule with the largest similarity, selecting the set number of rules. For example, the method described in claim 2, wherein the formula is adopted Calculate the similarity, where x, y F, F = ( F ₁ , F ₂ , ..., F _n ), i takes the value [1, n], and F ₀ ~ F _n are presets in the rule base to describe various types of advertising attributes. The set of F ₀ ~ F _{n is} used to form F _i , and j is the component contained in F _i . The method of claim 3, wherein the at least one advertisement corresponding to the rule-constrained scene is filtered according to the obtained at least one rule, comprising: obtaining, by using an advertisement search engine, based on the obtained at least one rule Corresponding alternative advertisements; using the formula H _result ( x , y )= e ^βS × H _similarity ( x , y ), calculating the probability campaign score of at least one rule obtained; according to the probability of the campaign rankings from large to small The rules are sorted, and starting from the rule with the highest probability of winning the campaign, selecting a set number of rules; determining at least one candidate advertisement corresponding to the set number of rules as the final selected advertisement. The method of any one of claims 1-4, monitoring an operation behavior generated by the user for the delivery of the at least one advertisement, and converting the collected related parameters into corresponding rules to update the rule base. The method includes: extracting a newly generated rule according to the collected related parameters, and calculating an effect score S _new and a support degree N _{new of} the newly generated rule; If the newly generated rule does not exist in the rule base, and S _new and N _new are respectively greater than the corresponding set thresholds, the newly generated rule is added to the rule base; if the newly generated rule already exists in the rule base, the calculation of the newly generated rule and the rule base of the original stored rules combined effect of the score S _together and combined support of _the N, if S _bonding and N _together are greater than the corresponding threshold value is set, then S _together and _the N saved to the rules library; if N or S _{bonded together} than the corresponding threshold value is set, the rule is deleted from the newly generated rule base. The method of claim 5, wherein the formula is adopted Calculate the effect score S _{new of} the newly generated rule, and use the formula Calculating the support degree N _{new of} the newly generated rule, wherein , w _i preset expert weight coefficient; Is a normalized function; F _{stat is} used to represent newly generated rules, x F _stat , SetF is the set of F _stat vectors recorded in a certain period of time. The method of claim 5, wherein the formula is adopted And calculating a new generation rule stored in the rule base of the original combined effect of regular score S _{bonded together} and consolidated support degree N, where, [alpha] and the preset expansion factor β, S and N _{old Old} original saved for the rule Performance score and support. The method of any one of claims 1-4, wherein, according to the feature attribute set, at least two rules matching the feature attribute set are acquired in a preset rule base, according to a genetic algorithm Cross-variation of the at least two rules. An apparatus for improving data transmission speed of a website, comprising: an obtaining unit, configured to obtain an operation behavior when a user browses a website Corresponding feature attribute set, and acquiring at least one rule matching the feature attribute set in the preset rule base according to the feature attribute set; the first processing unit is configured to filter out the rule according to the obtained at least one rule At least one advertisement corresponding to the constrained scene, and delivering the at least one advertisement to the user; the second processing unit is configured to monitor an operation behavior of the user for the delivery of the at least one advertisement, and convert the collected related parameters into The corresponding rule updates the rule base. A system for improving the speed of data transmission of a website, comprising: a rule base for storing various rules for searching for advertisements; and an advertisement delivery management device for obtaining corresponding features according to an operation behavior when the user browses the website Attribute collection, according to the feature attribute set, obtaining at least one rule matching the feature attribute set in a preset rule base, and filtering at least one advertisement corresponding to the scene restricted by the rule according to the obtained at least one rule And delivering the at least one advertisement to the user, and monitoring an operation behavior generated by the user for the delivery of the at least one advertisement, and converting the collected related parameters into corresponding rules to update the rule base.