Full-link strategy combination determination method and device, electronic equipment and storage medium

Abstract

The invention provides a full-link strategy combination determination method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring a plurality of initial strategy combinations; for each initial strategy combination, putting the initial strategy combination into an experiment to obtain experiment data, and determining index data corresponding to the initial strategy combination according to the experiment data; and under the condition that the index data corresponding to any initial strategy combination does not conform to the preset index, executing the following iterative processing: for each initial strategy combination, calculating a fitness value according to the experimental data corresponding to the initial strategy combination, selecting the initial strategy combination with the highest corresponding fitness value as a to-be-deformed combination, performing deformation processing on the to-be-deformed combination to obtain a new iteration strategy combination, and continuing to put the iteration strategy combination into the experiment until an iteration strategy combination of which the corresponding index data accords with a preset index is generated; and determining the iteration strategy combination of which the corresponding index data accords with the preset index as a target strategy combination.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to a method, device, electronic device and storage medium for determining a full-link strategy combination. Background Art

[0002] In e-commerce platforms and video service platforms, AB experiments are often used to optimize user experience and increase revenue. Specifically, AB experiments are conducted on a single page or module to orthogonally verify the elements in each link of the entire transaction chain (such as home page banner ads, pop-up ads, panoramic marketing, cash register page selling goods, prices, payment methods, promotional copy, etc.), so as to obtain the best strategy for each link, and then combine the best strategies for each link together to obtain the target combination strategy. For example, AB experiments are conducted on the home page banner ads, pop-up ads, panoramic marketing, cash register page selling goods, prices, payment methods, promotional copy, etc., to obtain the corresponding best strategies, and then obtain the target combination strategy corresponding to the entire chain.

[0003] However, in the process of determining the target combination strategy through the above method, the experiments on each link are independent, lacking comprehensive optimization of the entire transaction chain from a global perspective, and unable to guarantee the overall effect of the target combination strategy. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a method, device, electronic device and storage medium for determining a full-link strategy combination, so as to solve the problem that in the current process of determining the target combination strategy, the experiments for each link are independent, and there is a lack of comprehensive optimization of the entire transaction link from a global perspective, and the overall effect of the target combination strategy cannot be guaranteed. The specific technical solution is as follows:

[0005] In a first aspect, the present application provides a method for determining a full-link strategy combination, including:

[0006] Acquire multiple initial strategy combinations, wherein each of the initial strategy combinations includes a link execution strategy corresponding to each link in the full link;

[0007] For each initial strategy combination, the initial strategy combination is put into an experiment to obtain corresponding experimental data, and the indicator data corresponding to the initial strategy combination is determined according to the experimental data;

[0008] In the case that the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators, the following iterative processing is performed: for each initial strategy combination, the fitness value is calculated according to the experimental data corresponding to the initial strategy combination, the initial strategy combination with the highest corresponding fitness value is selected as the combination to be deformed, the combination to be deformed is deformed to obtain a new iterative strategy combination, and the iterative strategy combination is continuously put into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated;

[0009] The iterative strategy combination whose corresponding indicator data meets the preset indicator is determined as the target strategy combination.

[0010] In a possible implementation, the step of transforming the combination to be transformed to obtain a new iterative strategy combination includes:

[0011] Selecting a combination to be crossed from a plurality of iterative strategy combinations in an iterative cycle in which the combination to be deformed is located;

[0012] Selecting an element to be crossed from the combination to be crossed;

[0013] The corresponding elements in the combination to be transformed are replaced by the elements to be crossed to obtain the iterative strategy combination.

[0014] In a possible implementation, the step of transforming the combination to be transformed to obtain a new iterative strategy combination includes:

[0015] Determining the element to be mutated from the combination to be deformed;

[0016] Determine, among all links corresponding to the full link, a target link corresponding to the element to be mutated;

[0017] Determine a corresponding target execution strategy among all link execution strategies corresponding to the target link;

[0018] The elements to be mutated in the combination to be transformed are replaced with the target execution strategy to obtain an iterative strategy combination.

[0019] In a possible implementation, the experimental data includes: scoring data corresponding to at least one scoring indicator, and the calculating of the fitness value according to the experimental data corresponding to the initial strategy combination includes:

[0020] Determine the weight corresponding to each of the scoring indicators;

[0021] Perform a weighted sum operation on the scoring data corresponding to all the scoring indicators according to the weight corresponding to each of the scoring indicators to obtain a comprehensive scoring result;

[0022] The comprehensive scoring result is determined as the fitness value corresponding to the initial strategy combination.

[0023] In a possible implementation, the method further includes:

[0024] According to the experimental data corresponding to all initial strategy combinations and all iterative strategy combinations, a multi-dimensional analysis is performed on each link in the full link to obtain a multi-dimensional analysis result;

[0025] Determining strategy adjustment information according to the fitness values ​​corresponding to all initial strategy combinations and all iterative strategy combinations and the multi-dimensional analysis results;

[0026] Each link in the full link is adjusted according to the strategy adjustment information.

[0027] In a possible implementation, the obtaining of multiple initial strategy combinations includes:

[0028] Obtain a strategy combination set and a preset screening condition, wherein the strategy combination set includes all strategy combinations corresponding to the full link;

[0029] Screening the strategy combinations in the strategy combination set based on the screening condition to obtain candidate combinations;

[0030] Determine a combination score for each candidate combination according to a preset heuristic criterion;

[0031] The candidate combinations whose combined scores are higher than a preset threshold are determined as the initial strategy combinations.

[0032] In a possible implementation, the obtaining of multiple initial strategy combinations includes:

[0033] Acquire a strategy combination set, wherein the strategy combination set includes all strategy combinations corresponding to the full link;

[0034] Grouping all the policy combinations in the policy combination set to obtain a plurality of policy combination subsets, wherein each of the policy combination subsets contains policy combinations with the same characteristics;

[0035] Performing parallel optimization processing on a plurality of the strategy combination subsets to obtain an optimal strategy combination in each of the strategy combination subsets;

[0036] Each of the optimal strategy combinations is used as an initial strategy combination.

[0037] In a second aspect, the present application provides a full-link strategy combination determination device, including:

[0038] An acquisition module, used to acquire multiple initial strategy combinations, wherein each of the initial strategy combinations contains a link execution strategy corresponding to each link in the full link;

[0039] An experiment module is used to, for each initial strategy combination, put the initial strategy combination into an experiment to obtain corresponding experimental data, and determine the indicator data corresponding to the initial strategy combination according to the experimental data;

[0040] The iterative processing module is used to perform the following iterative processing when the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators: for each initial strategy combination, calculate the fitness value according to the experimental data corresponding to the initial strategy combination, select the initial strategy combination with the highest corresponding fitness value as the combination to be deformed, perform deformation processing on the combination to be deformed to obtain a new iterative strategy combination, and continue to put the iterative strategy combination into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated;

[0041] The determination module is used to determine the iterative strategy combination whose corresponding indicator data meets the preset indicator as the target strategy combination.

[0042] In a possible implementation, the iterative processing module is specifically configured to:

[0043] Selecting a combination to be crossed from a plurality of iterative strategy combinations in an iterative cycle in which the combination to be deformed is located;

[0044] Selecting an element to be crossed from the combination to be crossed;

[0045] The corresponding elements in the combination to be transformed are replaced by the elements to be crossed to obtain the iterative strategy combination.

[0046] In a possible implementation, the iterative processing module is further configured to:

[0047] Determining the element to be mutated from the combination to be deformed;

[0048] Determine, among all links corresponding to the full link, a target link corresponding to the element to be mutated;

[0049] Determine a corresponding target execution strategy among all link execution strategies corresponding to the target link;

[0050] The elements to be mutated in the combination to be transformed are replaced with the target execution strategy to obtain an iterative strategy combination.

[0051] In a possible implementation, the experimental data includes: scoring data corresponding to at least one scoring indicator, and the iterative processing module is further used to:

[0052] Determine the weight corresponding to each of the scoring indicators;

[0053] Perform a weighted sum operation on the scoring data corresponding to all the scoring indicators according to the weight corresponding to each of the scoring indicators to obtain a comprehensive scoring result;

[0054] The comprehensive scoring result is determined as the fitness value corresponding to the initial strategy combination.

[0055] In a possible implementation manner, the device further includes a policy adjustment module, configured to:

[0056] According to the experimental data corresponding to all initial strategy combinations and all iterative strategy combinations, a multi-dimensional analysis is performed on each link in the full link to obtain a multi-dimensional analysis result;

[0057] Determining strategy adjustment information according to the fitness values ​​corresponding to all initial strategy combinations and all iterative strategy combinations and the multi-dimensional analysis results;

[0058] Each link in the full link is adjusted according to the strategy adjustment information.

[0059] In a possible implementation, the acquisition module is specifically configured to:

[0060] Obtain a strategy combination set and a preset screening condition, wherein the strategy combination set includes all strategy combinations corresponding to the full link;

[0061] Screening the strategy combinations in the strategy combination set based on the screening condition to obtain candidate combinations;

[0062] Determine a combination score for each candidate combination according to a preset heuristic criterion;

[0063] The candidate combinations whose combined scores are higher than a preset threshold are determined as the initial strategy combinations.

[0064] In a possible implementation manner, the acquisition module is further used to:

[0065] Acquire a strategy combination set, wherein the strategy combination set includes all strategy combinations corresponding to the full link;

[0066] Grouping all the policy combinations in the policy combination set to obtain a plurality of policy combination subsets, wherein each of the policy combination subsets contains policy combinations with the same characteristics;

[0067] Performing parallel optimization processing on a plurality of the strategy combination subsets to obtain an optimal strategy combination in each of the strategy combination subsets;

[0068] Each of the optimal strategy combinations is used as an initial strategy combination.

[0069] In a third aspect, an electronic device is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0070] Memory, used to store computer programs;

[0071] The processor is used to implement any method step described in the first aspect when executing a program stored in the memory.

[0072] In a fourth aspect, a computer-readable storage medium is provided, characterized in that a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any method step described in the first aspect is implemented.

[0073] In a fifth aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute any of the above-described methods for determining a full-link strategy combination.

[0074] Beneficial effects of the embodiments of the present application:

[0075] The embodiment of the present application provides a method, device, electronic device and storage medium for determining a full-link strategy combination. The present application first obtains multiple initial strategy combinations, wherein each of the initial strategy combinations contains a link execution strategy corresponding to each link in the full link, then, for each initial strategy combination, the initial strategy combination is put into experiment to obtain corresponding experimental data, and the index data corresponding to the initial strategy combination is determined according to the experimental data. When the index data corresponding to any of the initial strategy combinations does not meet the preset index, the following iterative processing is performed: for each initial strategy combination, the fitness value is calculated according to the experimental data corresponding to the initial strategy combination, the initial strategy combination with the highest corresponding fitness value is selected as the combination to be deformed, the combination to be deformed is deformed to obtain a new iterative strategy combination, and the iterative strategy combination is continuously put into experiment until an iterative strategy combination whose corresponding index data meets the preset index is generated, and finally, the iterative strategy combination whose corresponding index data meets the preset index is determined as the target strategy combination. The scheme of the present application designs and optimizes the whole full link experimentally, and obtains the target strategy combination by integrating the experimental data of each link, thereby ensuring the overall effect of the target strategy combination and ensuring that the overall performance of the full link is optimal.

[0076] Of course, implementing any product or method of the present application does not necessarily require achieving all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0078] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0079] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0080] Figure 1 A flowchart of a method for determining a full-link strategy combination provided in an embodiment of the present application;

[0081] Figure 2 A flowchart of another full-link strategy combination determination method provided in an embodiment of the present application;

[0082] Figure 3 A flowchart of an embodiment of obtaining multiple initial strategy combinations provided in an embodiment of the present application;

[0083] Figure 4 A flowchart of another embodiment of obtaining multiple initial strategy combinations provided in an embodiment of the present application;

[0084] Figure 5 A system flow diagram of a full-link strategy combination determination method provided in an embodiment of the present application;

[0085] Figure 6 A schematic diagram of the structure of a full-link strategy combination determination device provided in an embodiment of the present application;

[0086] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0087] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0088] The disclosure below provides many different embodiments or examples to implement different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0089] Figure 1 A flow chart of a method for determining a full-link strategy combination provided for an embodiment of the present application. This method can be applied to one or more electronic devices such as smart phones, laptops, desktop computers, portable computers, servers, etc. In addition, the execution subject of this method can be hardware or software. When the above-mentioned execution subject is hardware, the execution subject can be one or more of the above-mentioned electronic devices. For example, a single electronic device can execute this method, or multiple electronic devices can cooperate with each other to execute this method. When the above-mentioned execution subject is software, this method can be implemented as multiple software or software modules, or as a single software or software module. It is not specifically limited here.

[0090] like Figure 1 As shown, the method specifically includes:

[0091] Step 101: Acquire multiple initial strategy combinations, wherein each of the initial strategy combinations contains a link execution strategy corresponding to each link in the entire link.

[0092] In applications, the full link generally refers to the full link used by users of various network platforms. For example, the full transaction link of the video application platform may include three links: home page banner advertising, floating ads, and checkout counter product display. Each link corresponds to multiple different link execution strategies. For example, the home page banner includes three link execution strategies: A1, A2, and A3; the floating ads include three link execution strategies: B1, B2, and B3; and the checkout counter product display includes three link execution strategies: C1, C2, and C3.

[0093] In an embodiment of the present application, an initial strategy combination can be obtained by combining the link execution strategies of each link. For example, by combining the link execution strategies of the three links of home page banner, floating advertisement and checkout counter product display, multiple initial strategy combinations such as [A1, B1, C1], [A2, B2, C2], [A3, B3, C3] can be obtained.

[0094] Step 102: for each initial strategy combination, subject the initial strategy combination to an experiment to obtain corresponding experimental data, and determine the indicator data corresponding to the initial strategy combination according to the experimental data.

[0095] The above experiments are used to evaluate the impact of different variables on specific indicators (such as conversion rate), for example, AB experiments.

[0096] Experimental data, including user behavior (such as clicks, purchases, payments, etc.) and conversion effects (such as revenue, click-through rate, etc.).

[0097] Indicator data refers to the data obtained from the experiment on the indicators that users focus on in conversion effects. At the beginning of the experiment, users can set the indicators of focus based on actual needs, such as revenue and retention.

[0098] In the embodiment of the present application, each initial strategy combination is put into the experimental platform for experiment to obtain corresponding experimental data, and the indicator data is obtained from the experimental data.

[0099] As a specific example, the experimental process is as follows:

[0100] 1. Determine the experimental parameters:

[0101] Target pages or modules: Determine which pages or modules you want to conduct the experiment on, such as homepage banner ads, floating ads, and checkout product displays.

[0102] Experimental user group: Determine the user group that will participate in the experiment, which can be all platform users or a targeted portion of users.

[0103] Experiment time window: Determine the time range for the experiment. Generally, the experiment needs to last long enough to ensure the accuracy of the data.

[0104] 2. Assign users

[0105] Random Assignment: Users are randomly assigned to different strategy combinations to ensure the fairness of the experiment and the representativeness of the data.

[0106] AB testing principles: In order to obtain reliable experimental results, at least two or more strategy combinations are required for comparative experiments. In order to ensure the representativeness and fairness of the data, the display ratio of each strategy combination is usually relatively even.

[0107] 3. Execute the experiment

[0108] Display strategy combination: Display different strategy combinations to assigned users on the specified page or module. For example, one user may see the homepage Banner A1 and floating ad B1, while another user will see the homepage Banner A2 and floating ad B2.

[0109] Data collection: Real-time collection of user behavior data under different strategy combinations, including clicks, browsing, purchases, viewing time, etc.

[0110] Step 103. When the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators, perform the following iterative processing: for each initial strategy combination, calculate the fitness value according to the experimental data corresponding to the initial strategy combination, select the initial strategy combination with the highest corresponding fitness value as the combination to be deformed, deform the combination to be deformed to obtain a new iterative strategy combination, and continue to put the iterative strategy combination into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated.

[0111] Step 104: Determine the iterative strategy combination whose corresponding indicator data meets the preset indicator as the target strategy combination.

[0112] Preset indicators refer to the preset thresholds of the indicators that users focus on in conversion effects.

[0113] For ease of understanding, step 103 and step 104 are described in a unified manner as follows:

[0114] In an embodiment of the present application, when the indicator data corresponding to all initial strategy combinations do not meet the preset indicators, based on the idea of ​​genetic algorithm, a strategy combination with high fitness is selected as the parent generation by simulating natural selection for subsequent crossover and mutation of new strategy combinations, thereby gradually finding the optimal solution, that is, the target strategy combination that meets the preset indicators.

[0115] Specifically, the experimental data includes: scoring data corresponding to at least one scoring indicator, and calculating the fitness value according to the experimental data corresponding to the initial strategy combination may include the following steps:

[0116] Determine the weight corresponding to each of the scoring indicators, perform a weighted sum operation on the scoring data corresponding to all of the scoring indicators according to the weight corresponding to each of the scoring indicators, obtain a comprehensive scoring result, and determine the comprehensive scoring result as the fitness value corresponding to the initial strategy combination.

[0117] As a specific example, scoring metrics may include the following:

[0118] Revenue: The total amount of money users spend on the platform.

[0119] Click-through rate: How often users click on your ad.

[0120] Conversion rate: The percentage of users who purchase after clicking on an ad.

[0121] User retention rate: the proportion of users who continue to use the platform after a period of time.

[0122] Watch time: The total amount of time a user watches a video.

[0123] Correspondingly, the scoring data is the experimental data collected during the experiment:

[0124] Revenue: How much money the user spent in total under this strategy combination.

[0125] Click-through rate: The percentage of users who click on your ad when your ad is shown to them.

[0126] Conversion rate: What percentage of people who click on your ad actually purchase something.

[0127] User retention rate: How many users are still using the platform after one month.

[0128] Watch time: the total amount of time users watch videos on the platform.

[0129] Then, the above scoring data is converted to a unified scoring standard, so that all data can be compared at the same level. For example, for "income", the strategy combination brings different data from 1,000 to 2,000 yuan. 1,000 yuan can be matched with 0 points, 2,000 yuan can be matched with 100 points, and other amounts in between can be matched in proportion.

[0130] Furthermore, weights can be assigned to different scoring indicators according to their importance: for example, if the user believes that "income" is the most important, a higher weight (such as 40%) can be set for it.

[0131] "Click-through rate" is also relatively important, and a 30% weight can be set. "Conversion rate" is relatively unimportant, and a 20% weight can be set. "User retention rate" can be set at 10%.

[0132] Finally, the total score (i.e., fitness value) of each strategy combination is calculated by weighted summation:

[0133] For example, a strategy combination scored 80 points in "revenue", 70 points in "click-through rate", 50 points in "conversion rate", and 60 points in "user retention rate". Combine these scores and their corresponding weights to get the total score (fitness value):

[0134] Total score = revenue score × 40% + click-through rate score × 30% + conversion rate score × 20% + user retention rate score × 10%.

[0135] Therefore, for the above strategy combination, the total score calculation result is as follows: Total score = 80×0.4+70×0.3+50×0.2+60×0.1.

[0136] In this way, the overall performance of the current strategy combination can be understood through the fitness value.

[0137] In addition, in one embodiment, the specific implementation of performing deformation processing on the combination to be deformed to obtain a new iterative strategy combination may include the following steps:

[0138] Select a combination to be crossed from among multiple iteration strategy combinations in the iteration cycle of the combination to be deformed; select an element to be crossed from the combination to be crossed; replace the corresponding element in the combination to be deformed with the element to be crossed to obtain the iteration strategy combination.

[0139] In this implementation, for each iteration cycle, the combination to be transformed and the randomly selected combination to be crossed in the iteration cycle are used as the parent strategy combination, and element crossover is performed to obtain the iteration strategy combination.

[0140] For example, the parent combinations are [A2, B2, C2] and [A1, B1, C1], and the elements of the two are crossed to generate a new iteration strategy combination [A2, B1, C2].

[0141] In another embodiment, the specific implementation of performing deformation processing on the combination to be deformed to obtain a new iterative strategy combination may include the following steps:

[0142] Determine the element to be mutated from the combination to be transformed; determine the target link corresponding to the element to be mutated among all links corresponding to the full link; determine the corresponding target execution strategy among all link execution strategies corresponding to the target link; replace the element to be mutated in the combination to be transformed with the target execution strategy to obtain an iterative strategy combination.

[0143] In this implementation, a target link and the target execution strategy corresponding to the target link can be randomly selected, and the elements to be mutated in the combination to be transformed are replaced with the target execution strategy to obtain an iterative strategy combination. For example, [A2, B1, C2] is mutated to generate [A2, B1, C3]. In this way, the diversity of the combination can be increased by randomly changing the genes in the combination.

[0144] When there is an initial strategy combination whose corresponding indicator data meets the preset indicators, it can be put into use as the target strategy combination, and wait for the next cycle (such as one day, one week or one month as a cycle) to execute steps 101 to 104 to adjust the strategy combination executed in the platform in real time, thereby ensuring the effectiveness of the platform.

[0145] In an embodiment of the present application, first, a plurality of initial strategy combinations are obtained, wherein each of the initial strategy combinations contains a link execution strategy corresponding to each link in the full link, and then, for each initial strategy combination, the initial strategy combination is put into an experiment to obtain corresponding experimental data, and the index data corresponding to the initial strategy combination is determined according to the experimental data. When the index data corresponding to any of the initial strategy combinations does not meet the preset index, the following iterative processing is performed: for each initial strategy combination, the fitness value is calculated according to the experimental data corresponding to the initial strategy combination, the initial strategy combination with the highest corresponding fitness value is selected as the combination to be deformed, the combination to be deformed is deformed to obtain a new iterative strategy combination, and the iterative strategy combination is continuously put into the experiment until an iterative strategy combination whose corresponding index data meets the preset index is generated, and finally, the iterative strategy combination whose corresponding index data meets the preset index is determined as the target strategy combination. The scheme of the present application, by designing and optimizing the whole full link experiment, obtains the target strategy combination by integrating the experimental data of each link, thereby ensuring the overall effect of the target strategy combination and ensuring that the overall performance of the full link is optimal.

[0146] See also Figure 2 , is a flow chart of another embodiment of a method for determining a full-link strategy combination provided by an embodiment of the present application. Figure 2 As shown, the process may include the following steps:

[0147] Step 201: Perform a multi-dimensional analysis on each link in the full link according to the experimental data corresponding to all initial strategy combinations and all iterative strategy combinations to obtain a multi-dimensional analysis result;

[0148] Step 202: Determine strategy adjustment information according to the fitness values ​​corresponding to all initial strategy combinations and all iterative strategy combinations and the multi-dimensional analysis results;

[0149] Step 203: Adjust each link in the full link according to the strategy adjustment information.

[0150] For ease of understanding, steps 201 to 203 are described in a unified manner as follows:

[0151] In an embodiment of the present application, the analysis dimensions can be set by the user according to actual needs, for example, click-through rate dimension, conversion rate dimension, revenue dimension, user retention rate dimension, etc.

[0152] As an example, the analysis process is as follows:

[0153] 1. Data cleaning and preprocessing

[0154] Data cleaning: Remove erroneous, missing, and inconsistent data to ensure data quality.

[0155] Data preprocessing: Perform preparatory work such as format conversion and normalization to make the data suitable for subsequent analysis.

[0156] 2. Statistical analysis of indicators in each dimension

[0157] Basic statistics: calculate the mean, median, standard deviation, extreme value, etc. of each indicator.

[0158] Click-through rate: Statistics on the mean and distribution of click-through rate for each strategy combination.

[0159] Conversion rate: Statistics on the average conversion rate and its fluctuations.

[0160] Income: Statistics on the average income and total income under each strategy combination.

[0161] User retention rate: Statistics show the proportion of users who continue to use the platform after a period of time.

[0162] Multi-dimensional analysis results:

[0163] Click-through rate: The click-through rate of different strategy combinations, identifying which strategy combination attracts the most users to click.

[0164] Conversion rate: Evaluate the impact of strategy combinations on user purchasing behavior and find the strategy with the highest conversion rate.

[0165] Revenue: Analyze the contribution of each strategy combination to the platform revenue and find the strategy that is most likely to bring high revenue.

[0166] User retention rate: Evaluate the effect of strategy combinations on long-term user retention and find out which strategy is most effective in retaining users.

[0167] Accordingly, the policy adjustment information may include the following situations:

[0168] Case 1: The fitness value of the overall strategy combination is low:

[0169] Reason: In this case, the overall performance of the strategy portfolio is poor, and some key indicators may not perform well.

[0170] Adjustment suggestions: Check various indicators to determine which specific indicators (revenue, click-through rate, etc.) are lagging behind. If the revenue is low, try to adjust the pricing strategy or promote more attractive products to increase high-quality content to attract user consumption; if the click-through rate is low, improve the ad creative, ad location or ad copy to optimize the page layout and improve the visibility and attractiveness of the ad.

[0171] Case 2: One indicator performs well, while other indicators perform generally well:

[0172] Cause: One possibility is that the strategy optimizes one aspect more, but ignores other aspects.

[0173] Adjustment suggestions: Balance optimization, continue to maintain the advantages in strengths, and improve performance in other aspects. If the click-through rate is high but the conversion rate is low, check the consistency of the ad content and the landing page, and ensure that the page after the click is attractive to users, so as to increase the clear conversion path and guide users to complete the purchase; if the revenue is high but the user retention rate is low, provide more value and discounts to increase users' willingness to continue using, so as to improve the user experience and ensure that they have a good experience on the platform.

[0174] Case 3: The fitness value is high, but there is room for further optimization:

[0175] Reason: The strategy performs well, but could still be optimized with some minor tweaks.

[0176] Adjustment suggestions: Refine the analysis, check the segmented data of each indicator, and find possible optimization points. If the overall fitness value is high, continue to improve and A / B test to verify whether the new strategy combination can further improve performance, so as to improve the scores of individual small differences, and improve through fine-tuning such as optimizing user paths, enhancing content and service quality.

[0177] In the embodiment of the present application, the corresponding links in the whole link can be adjusted through the fitness values ​​and multi-dimensional analysis results of each strategy combination, so that the overall strategy combination of the whole link can better meet the user needs.

[0178] See also Figure 3 , is a flowchart of an embodiment of obtaining multiple initial strategy combinations provided by the embodiment of the present application. Figure 3 As shown, the process may include the following steps:

[0179] Step 301: Obtain a policy combination set and a preset screening condition, wherein the policy combination set includes all policy combinations corresponding to the full link;

[0180] Step 302: Screen the policy combinations in the policy combination set based on the screening condition to obtain candidate combinations;

[0181] Step 303: Determine a combination score for each candidate combination according to a preset heuristic standard;

[0182] Step 304: Determine the candidate combination whose combination score is higher than the preset threshold as the initial strategy combination.

[0183] For ease of understanding, steps 301 to 304 are described in a unified manner as follows:

[0184] The above screening conditions can be set by the user according to actual needs. For example, a minimum click-through rate and conversion rate can be set, and only strategy combinations that meet these standards will be retained. In this way, those combinations that perform significantly poorly can be quickly eliminated.

[0185] Heuristic criteria can be standards set based on existing data and experience to predict which combinations are most likely to perform well. For example, a comprehensive score is calculated by weighting indicators such as click-through rate and conversion rate, and strategy combinations with scores higher than a preset threshold are retained.

[0186] In the embodiment of the present application, first, the combinations with obviously poor performance are screened out through screening conditions to obtain candidate combinations that may perform well. Then, heuristic search is performed based on heuristic criteria to find and optimize strategies in a more targeted manner, avoiding wasting time and resources in unlikely areas, thereby obtaining an initial strategy combination.

[0187] See also Figure 4 , is another flow chart of an embodiment of the present application for obtaining multiple initial strategy combinations. Figure 4 As shown, the process may include the following steps:

[0188] Step 401: Obtain a policy combination set, wherein the policy combination set includes all policy combinations corresponding to the full link;

[0189] Step 402: Group all the policy combinations in the policy combination set to obtain multiple policy combination subsets, wherein each of the policy combination subsets contains policy combinations with the same characteristics;

[0190] Step 403: performing parallel optimization processing on the plurality of strategy combination subsets to obtain the optimal strategy combination in each strategy combination subset;

[0191] Step 404: Take each of the optimal strategy combinations as an initial strategy combination.

[0192] For ease of understanding, steps 401 to 404 are described in a unified manner as follows:

[0193] In the embodiment of the present application, the strategy combinations can be divided into different groups according to the characteristics of the strategies. For example, similar advertisements, similar user groups, etc. Then, a step-by-step optimization is performed layer by layer, that is, multiple groups are optimized in parallel first to find the best strategy combination in each group. Then, the best strategy combination of each group is used as the initial strategy combination for further overall optimization.

[0194] In the embodiments of the present application, through grouping and hierarchical optimization, we can process a large number of strategy combinations in a more organized manner, reducing computational complexity and time consumption.

[0195] Optionally, the embodiment of the present application also provides a system flow diagram of the method for determining a full-link strategy combination, such as Figure 5 As shown, the specific steps are as follows.

[0196] a. User->Policy Management Module: Define and update policies

[0197] User-defined and updated policies: Users enter and configure policy combinations for different pages and modules in the policy management module. For example, users may define different policies for home page banner ads, floating ads, and cashier product displays.

[0198] b. Policy management module->Global policy combination generator: provides policy configuration

[0199] The policy management module provides policy configuration: the policy management module transmits the user-defined policy combination to the global policy combination generator as the basis for combination generation.

[0200] c. Global strategy combination generator -> automated experiment management: provides global strategy combinations

[0201] The global strategy combination generator provides a global strategy combination: The global strategy combination generator uses a genetic algorithm to generate an initial strategy combination, which includes the following steps:

[0202] Initial population generation: Randomly generate several initial strategy combinations, such as home page banner (A1, A2, A3), floating ads (B1, B2, B3), and cash register product displays (C1, C2, C3).

[0203] Fitness calculation: Calculate the fitness value of each strategy combination based on the experimental results (such as revenue, click-through rate, etc.).

[0204] Natural selection: Select strategy combinations with high fitness values ​​as parents.

[0205] Crossover and mutation: Generate new strategy combinations by simulating natural selection and crossover mutation.

[0206] Iterative optimization: Repeat the above process continuously to gradually optimize the strategy combination.

[0207] d. Automated experiment management -> Experimental data storage: record experimental data

[0208] Automated experiment management records experimental data: The automated experiment management module configures and executes strategy experiments, and records the data collected during the experiment (such as user clicks, conversion rates, etc.) into the experimental data storage.

[0209] e. Automated experiment management -> transaction data storage: record transaction data

[0210] Automated experiment management records transaction data:

[0211] The automated experiment management module records user transaction behaviors (such as purchases, payments, etc.) into the transaction data storage.

[0212] f. Automated experiment management -> Big data analysis module: Sending experimental result data

[0213] Automated experiment management sends experimental result data: The automated experiment management module sends the collected experimental result data to the big data analysis module for analysis and processing.

[0214] g. Big data analysis module->Experimental data storage: storage and analysis data

[0215] Big data analysis module stores analysis data: The big data analysis module performs multi-dimensional analysis on the experimental data and stores the analysis results in the experimental data storage.

[0216] h. Big data analysis module -> strategy management module: provide strategy adjustment suggestions

[0217] The big data analysis module provides strategy adjustment suggestions: The big data analysis module calculates the fitness value of each strategy combination based on the experimental data and analysis results, and provides strategy adjustment suggestions to the strategy management module.

[0218] In this way, the automated experiment management system can be used to realize intelligent management of experiment configuration, execution and monitoring, and support real-time data analysis and dynamic strategy adjustment, thereby improving experiment management efficiency, reducing labor costs and optimizing experiment results.

[0219] Based on the same technical concept, the embodiment of the present application also provides a full-link strategy combination determination device, such as Figure 6 The device comprises:

[0220] An acquisition module 61 is used to acquire multiple initial strategy combinations, wherein each of the initial strategy combinations contains a link execution strategy corresponding to each link in the full link;

[0221] The experiment module 62 is used for subjecting each initial strategy combination to an experiment to obtain corresponding experimental data, and determining the indicator data corresponding to the initial strategy combination according to the experimental data;

[0222] The iterative processing module 63 is used to perform the following iterative processing when the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators: for each initial strategy combination, calculate the fitness value according to the experimental data corresponding to the initial strategy combination, select the initial strategy combination with the highest corresponding fitness value as the combination to be deformed, perform deformation processing on the combination to be deformed to obtain a new iterative strategy combination, and continue to put the iterative strategy combination into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated;

[0223] The determination module 64 is used to determine the iterative strategy combination whose corresponding indicator data meets the preset indicator as the target strategy combination.

[0224] In a possible implementation, the iterative processing module is specifically configured to:

[0225] Selecting a combination to be crossed from a plurality of iterative strategy combinations in an iterative cycle in which the combination to be deformed is located;

[0226] Selecting an element to be crossed from the combination to be crossed;

[0227] The corresponding elements in the combination to be transformed are replaced by the elements to be crossed to obtain the iterative strategy combination.

[0228] In a possible implementation, the iterative processing module is further configured to:

[0229] Determining the element to be mutated from the combination to be deformed;

[0230] Determine, among all links corresponding to the full link, a target link corresponding to the element to be mutated;

[0231] Determine a corresponding target execution strategy among all link execution strategies corresponding to the target link;

[0232] The elements to be mutated in the combination to be transformed are replaced with the target execution strategy to obtain an iterative strategy combination.

[0233] In a possible implementation, the experimental data includes: scoring data corresponding to at least one scoring indicator, and the iterative processing module is further used to:

[0234] Determine the weight corresponding to each of the scoring indicators;

[0235] Perform a weighted sum operation on the scoring data corresponding to all the scoring indicators according to the weight corresponding to each of the scoring indicators to obtain a comprehensive scoring result;

[0236] The comprehensive scoring result is determined as the fitness value corresponding to the initial strategy combination.

[0237] In a possible implementation manner, the device further includes a policy adjustment module, configured to:

[0238] According to the experimental data corresponding to all initial strategy combinations and all iterative strategy combinations, a multi-dimensional analysis is performed on each link in the full link to obtain a multi-dimensional analysis result;

[0239] Determining strategy adjustment information according to the fitness values ​​corresponding to all initial strategy combinations and all iterative strategy combinations and the multi-dimensional analysis results;

[0240] Each link in the full link is adjusted according to the strategy adjustment information.

[0241] In a possible implementation, the acquisition module is specifically configured to:

[0242] Obtain a strategy combination set and a preset screening condition, wherein the strategy combination set includes all strategy combinations corresponding to the full link;

[0243] Screening the strategy combinations in the strategy combination set based on the screening condition to obtain candidate combinations;

[0244] Determine a combination score for each candidate combination according to a preset heuristic criterion;

[0245] The candidate combinations whose combined scores are higher than a preset threshold are determined as the initial strategy combinations.

[0246] In a possible implementation manner, the acquisition module is further used to:

[0247] Acquire a strategy combination set, wherein the strategy combination set includes all strategy combinations corresponding to the full link;

[0248] Grouping all the policy combinations in the policy combination set to obtain a plurality of policy combination subsets, wherein each of the policy combination subsets contains policy combinations with the same characteristics;

[0249] Performing parallel optimization processing on a plurality of the strategy combination subsets to obtain an optimal strategy combination in each of the strategy combination subsets;

[0250] Each of the optimal strategy combinations is used as an initial strategy combination.

[0251] Based on the same technical concept, the embodiment of the present application also provides an electronic device, such as Figure 7 As shown, it includes a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0252] Memory 113, used for storing computer programs;

[0253] The processor 111 is used to execute the program stored in the memory 113 to implement the following steps:

[0254] Acquire multiple initial strategy combinations, wherein each of the initial strategy combinations includes a link execution strategy corresponding to each link in the full link;

[0255] For each initial strategy combination, the initial strategy combination is put into an experiment to obtain corresponding experimental data, and the indicator data corresponding to the initial strategy combination is determined according to the experimental data;

[0256] In the case that the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators, the following iterative processing is performed: for each initial strategy combination, the fitness value is calculated according to the experimental data corresponding to the initial strategy combination, the initial strategy combination with the highest corresponding fitness value is selected as the combination to be deformed, the combination to be deformed is deformed to obtain a new iterative strategy combination, and the iterative strategy combination is continuously put into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated;

[0257] The iterative strategy combination whose corresponding indicator data meets the preset indicator is determined as the target strategy combination.

[0258] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0259] The communication interface is used for communication between the above electronic device and other devices.

[0260] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0261] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0262] In another embodiment provided in the present application, a computer-readable storage medium is also provided, in which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned full-link strategy combination determination methods are implemented.

[0263] In another embodiment provided in the present application, a computer program product including instructions is also provided, which, when executed on a computer, enables the computer to execute any of the full-link strategy combination determination methods in the above embodiments.

[0264] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0265] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0266] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used herein may also be meant to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The method steps, processes, and operations described herein are not interpreted as necessarily requiring them to be performed in the specific order described or illustrated, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps may be used.

[0267] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for determining a full-link strategy combination, characterized in that: The method comprises: Acquire multiple initial strategy combinations, wherein each of the initial strategy combinations includes a link execution strategy corresponding to each link in the full link; For each initial strategy combination, the initial strategy combination is put into an experiment to obtain corresponding experimental data, and the indicator data corresponding to the initial strategy combination is determined according to the experimental data; In the case that the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators, the following iterative processing is performed: for each initial strategy combination, the fitness value is calculated according to the experimental data corresponding to the initial strategy combination, the initial strategy combination with the highest corresponding fitness value is selected as the combination to be deformed, the combination to be deformed is deformed to obtain a new iterative strategy combination, and the iterative strategy combination is continuously put into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated; The iterative strategy combination whose corresponding indicator data meets the preset indicator is determined as the target strategy combination.

2. The method according to claim 1, characterized in that The step of performing deformation processing on the combination to be deformed to obtain a new iterative strategy combination includes: Selecting a combination to be crossed from a plurality of iterative strategy combinations in an iterative cycle in which the combination to be deformed is located; Selecting an element to be crossed from the combination to be crossed; The corresponding elements in the combination to be transformed are replaced by the elements to be crossed to obtain the iterative strategy combination.

3. The method according to claim 1, characterized in that: The step of performing deformation processing on the combination to be deformed to obtain a new iterative strategy combination includes: Determining the element to be mutated from the combination to be deformed; Determine, among all links corresponding to the full link, a target link corresponding to the element to be mutated; Determine a corresponding target execution strategy among all link execution strategies corresponding to the target link; The elements to be mutated in the combination to be transformed are replaced with the target execution strategy to obtain an iterative strategy combination.

4. The method according to claim 1, characterized in that: The experimental data includes: scoring data corresponding to at least one scoring indicator, and the calculating of the fitness value according to the experimental data corresponding to the initial strategy combination includes: Determine the weight corresponding to each of the scoring indicators; Perform a weighted sum operation on the scoring data corresponding to all the scoring indicators according to the weight corresponding to each of the scoring indicators to obtain a comprehensive scoring result; The comprehensive scoring result is determined as the fitness value corresponding to the initial strategy combination.

5. The method according to claim 1, characterized in that The method further comprises: According to the experimental data corresponding to all initial strategy combinations and all iterative strategy combinations, a multi-dimensional analysis is performed on each link in the full link to obtain a multi-dimensional analysis result; Determining strategy adjustment information according to the fitness values ​​corresponding to all initial strategy combinations and all iterative strategy combinations and the multi-dimensional analysis results; Each link in the full link is adjusted according to the strategy adjustment information.

6. The method according to claim 1, characterized in that The obtaining of multiple initial strategy combinations includes: Obtain a strategy combination set and a preset screening condition, wherein the strategy combination set includes all strategy combinations corresponding to the full link; Screening the strategy combinations in the strategy combination set based on the screening condition to obtain candidate combinations; Determine a combination score for each candidate combination according to a preset heuristic criterion; The candidate combinations whose combined scores are higher than a preset threshold are determined as the initial strategy combinations.

7. The method according to claim 1, characterized in that The obtaining of multiple initial strategy combinations includes: Acquire a strategy combination set, wherein the strategy combination set includes all strategy combinations corresponding to the full link; Grouping all the policy combinations in the policy combination set to obtain a plurality of policy combination subsets, wherein each of the policy combination subsets contains policy combinations with the same characteristics; Performing parallel optimization processing on a plurality of the strategy combination subsets to obtain an optimal strategy combination in each of the strategy combination subsets; Each of the optimal strategy combinations is used as an initial strategy combination.

8. A device for determining a full-link strategy combination, characterized in that: The device comprises: An acquisition module, used to acquire multiple initial strategy combinations, wherein each of the initial strategy combinations contains a link execution strategy corresponding to each link in the full link; An experiment module is used to, for each initial strategy combination, put the initial strategy combination into an experiment to obtain corresponding experimental data, and determine the indicator data corresponding to the initial strategy combination according to the experimental data; The iterative processing module is used to perform the following iterative processing when the indicator data corresponding to any of the initial strategy combinations do not meet the preset indicators: for each initial strategy combination, calculate the fitness value according to the experimental data corresponding to the initial strategy combination, select the initial strategy combination with the highest corresponding fitness value as the combination to be deformed, perform deformation processing on the combination to be deformed to obtain a new iterative strategy combination, and continue to put the iterative strategy combination into the experiment until an iterative strategy combination whose corresponding indicator data meets the preset indicators is generated; The determination module is used to determine the iterative strategy combination whose corresponding indicator data meets the preset indicator as the target strategy combination.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, used to implement the full-link strategy combination determination method described in any one of claims 1-7 when executing a program stored in a memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for determining a full-link strategy combination according to any one of claims 1 to 7 is implemented.

Images