CN114266606B - Short video advertisement putting optimization evaluation method and system - Google Patents
Short video advertisement putting optimization evaluation method and system Download PDFInfo
- Publication number
- CN114266606B CN114266606B CN202210203897.2A CN202210203897A CN114266606B CN 114266606 B CN114266606 B CN 114266606B CN 202210203897 A CN202210203897 A CN 202210203897A CN 114266606 B CN114266606 B CN 114266606B
- Authority
- CN
- China
- Prior art keywords
- level
- delivery
- short video
- social
- putting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention discloses a short video advertisement putting optimization evaluation method and a system, comprising the following steps: step S1, acquiring a social network representing the hierarchical social relationship of the target user according to the social communication data, and clustering and quantizing the social network into a plurality of hierarchical topologies; and step S2, carrying out similarity matching on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and taking the short video advertisements to be launched with high similarity matching degree as the quasi-launched short video advertisements which are launched to the target users in an increment launching level in increments according to launching priorities. The method constructs the relevance of the delivery increment between the delivery levels with the high delivery priority and the low delivery priority based on the advertisement conversion rate of the delivery level with the high delivery priority, and then adjusts the delivery amount of the delivery level with the low delivery priority based on the relevance so as to realize the optimization of the advertisement conversion rate and the delivery expenditure.
Description
Technical Field
The invention relates to the technical field of advertisement putting, in particular to a short video advertisement putting optimization evaluation method and system.
Background
In recent years, online advertising has gained increasing weight throughout the advertising industry. The real-time bidding advertisement in the online advertisement has a high proportion year by year due to the good conversion effect. The DSP (Demand-Side Platform) is used as a Demand party Platform, and bid attempt is carried out on each exposure through an advertisement trading Platform. For each bidding request, the DSP tries to map users browsing a media website and App to users which can be identified by the DSP according to Cookie Mapping or device information, and then performs flow screening, click rate/conversion rate estimation and the like according to user figures mined from user historical behaviors by the DSP so as to contribute to maximization of the ROI.
The prior art has the problems that the acquired data are relatively simple, only the user data in the station is known, and the access behaviors of the user on other websites are unknown; on the other hand, after the user data is dense enough, the descriptive label attribute of the user shows higher stability, but the above method has the defect of poor generalization capability, thereby reducing the adaptability of the delivery evaluation.
Disclosure of Invention
The invention aims to provide a short video advertisement putting optimization evaluation method and a short video advertisement putting optimization evaluation system, which aim to solve the technical problem that the adaptability of putting evaluation is reduced due to the defect of poor generalization capability in the prior art.
In order to solve the technical problems, the invention specifically provides the following technical scheme:
a short video advertisement putting optimization evaluation method comprises the following steps:
step S1, acquiring a social network representing the hierarchical social relationship of the target user according to the social communication data, clustering and quantizing the social network into a plurality of hierarchical topologies, and marking the plurality of hierarchical topologies as incremental release levels with release priorities;
step S2, similarity matching is carried out on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and the short video advertisements to be launched with high similarity matching degree are used as quasi-launched short video advertisements which are launched to the increment launching level of the target users in an increment launching mode according to launching priorities;
step S3, counting the advertisement conversion rate of each delivery level in the incremental delivery levels according to the delivery priority, constructing the delivery incremental relevance between the delivery levels with high delivery priority and low delivery priority based on the advertisement conversion rate of the delivery level with high delivery priority, and adjusting the delivery amount of the delivery level with low delivery priority based on the relevance until the delivery of the quasi-delivery short video advertisement in all the delivery levels is completed according to the delivery amount, so as to realize the incremental adjustment of the delivery amount of the short video advertisement to achieve the optimization of the advertisement conversion rate and the delivery expenditure.
As a preferred aspect of the present invention, the obtaining a social networking hierarchy network representing a hierarchical social relationship of a target user according to social communication data includes:
setting a communication data volume threshold value representing a normalized social relationship, extracting a communication object of which the social communication data volume with a target user exceeds the communication data volume threshold value from the social communication data, and constructing the communication object as a level 1 topological node representing the level 1 social relationship of the target user;
extracting communication objects with the social communication data quantity of the 1 st to 5 th level topology nodes exceeding a communication data quantity threshold value from the social communication data in sequence, and respectively constructing the communication objects as 2 nd to 6 th level topology nodes representing the 2 nd to 6 th level social relations of the target user;
quantizing the target user into 0 th-level topological nodes, sequentially connecting 0 th-6 th-level topological nodes to obtain the social network, and taking the connection relation of the topological nodes as the edge relation of the social network.
As a preferred aspect of the present invention, the quantizing the clustering of the social hierarchy network into multiple hierarchy topologies includes:
clustering operation is carried out on the social hierarchical network based on modularity to obtain a plurality of topological communities, the topological communities where target users are located in the topological communities are used as the level 1 topology, and the rest topological communities are arranged in a reverse order according to the degree of a margin coefficient of the level 1 topology to be used as the level 1 topology
The hierarchical topology, wherein m is characterized as the total number of topological communities, and j is a metering constant;
preferably, the second step is sequentially calculated based on the amount of the side relation coefficient
The social affinity of the hierarchical topology to the level 1 topology includes:
in turn will be
The edge relation data quantity of the level topology and the level 1 topology is normalized to obtain the level one
The social affinity of the hierarchical topology and the 1 st hierarchical topology is calculated according to the formula:
in the formula (I), the compound is shown in the specification,
characterized by a social affinity of the jth hierarchical topology to the level 1 topology,
the number of edge relationships represented by the jth level topology and the level 1 topology.
As a preferred aspect of the present invention, the marking of the multiple levels of topology as the incremental delivery level with the delivery priority includes:
in turn will be
The releasing priority of the hierarchical topology is set to be k level, the hierarchical topology with the k level releasing priority is marked as the kth releasing level with the k level releasing priority, and all releasing levels are arranged according to the priority to form the incremental releasing level.
As a preferred scheme of the present invention, the similarity matching between the consumption attributes of the target users and the product attributes of the short video advertisements to be delivered in sequence includes:
quantizing the consumption attribute of the target user from a semantic form into a vector form by using an NLP tool word2vec to obtain a consumption attribute vector, and quantizing the product attribute of the short video advertisement to be delivered from the semantic form into the vector form by using the NLP tool word2vec to obtain a product attribute vector;
establishing a similarity matching formula of the consumption attribute of the target user and the product attribute of the short video advertisement to be launched based on the Euclidean distance of the consumption attribute vector and the product attribute vector, wherein the similarity matching formula is as follows:
in the formula (I), the compound is shown in the specification,
characterized by the similarity of the product attribute of the r-th short video advertisement to be delivered and the consumption attribute of the target user,
、
respectively representing the product attribute vector and the consumption attribute vector, wherein r is a metering constant;
setting a similarity matching threshold, and comparing the similarity of the product attribute of the short video advertisement to be launched and the consumption attribute of the target user with the similarity matching threshold in sequence, wherein,
if the similarity is
If the similarity matching threshold is exceeded, the similarity is compared
The corresponding r-th short video advertisement to be delivered is taken as a standard for delivering the short video advertisement;
if the similarity is
If the similarity matching threshold is not exceeded, the similarity is compared
And the corresponding r-th short video advertisement to be delivered is not taken as the short video advertisement to be delivered.
As a preferred scheme of the present invention, counting the advertisement conversion rate of each delivery level in incremental delivery levels according to the delivery priority includes:
counting the ratio of the number of users purchasing products in the r-th quasi-delivery short video advertisement in the k-th delivery layer of the incremental delivery layer to all the users in the k-th delivery layer in turn as the advertisement conversion rate of the r-th quasi-delivery short video advertisement in the k-th delivery layer, wherein the calculation formula of the advertisement conversion rate is as follows:
in the formula (I), the compound is shown in the specification,
characterized by the ad conversion rate of the r-th short video advertisement to be delivered in the k-th delivery level,
characterized by the number of users in the kth placement level who purchased the product in the r-th quasi-placement short video advertisement,
the characteristics are all the users in the k-th release level.
As a preferred aspect of the present invention, the constructing a delivery increment association between delivery levels with a high delivery priority and a low delivery priority based on an advertisement conversion rate of a delivery level with a high delivery priority includes:
obtaining the advertisement conversion rate of the delivery layer with high delivery priority, setting an attenuation coefficient for the advertisement conversion rate of the delivery layer with high delivery priority, attenuating the advertisement conversion rate based on the attenuation coefficient to obtain a correlation coefficient representing the relevance of the delivery increment, wherein,
the attenuation coefficient is obtained by constructing social affinity corresponding to the release level so as to realize attenuation according to the social affinity and improve attenuation scientificity, and the calculation formula of the attenuation coefficient is as follows:
in the formula (I), the compound is shown in the specification,
characterized by attenuation coefficients from the k-th level of delivery to the k + 1-th level of delivery,
the social affinity is characterized by the social affinity of the k +1 th release level and the 1 st release level;
the calculation formula of the correlation coefficient is as follows:
in the formula (I), the compound is shown in the specification,
the characteristic is that the r-th short video advertisement to be delivered is delivered with the association coefficient of the increment in the k-th delivery level and the k + 1-th delivery level;
the release increment is characterized by the increasing rate of the release amount of the k + 1-level release level on the basis of the release amount of the k-level release level.
As a preferred aspect of the present invention, the adjusting the placement amount of the placement level having the low placement priority based on the relevance includes:
according to the correlation coefficient, the first
The amount of the level of putting, the second
The calculation formula of the putting amount of the level putting level is as follows:
in the formula (I), the compound is shown in the specification,
、
respectively representing the putting quantities of the r-th quasi-put short video advertisement in the k +1 th and k-th putting levels;
setting the throwing amount of the 1 st level throwing level according to the order
And calculating the putting amount of the r-th quasi-delivery short video advertisement in the 1 st to m-th delivery levels by using a calculation formula of the putting amount of the level delivery levels.
As a preferred scheme of the present invention, the placement amount of the level 1 placement layer is in a direct proportion relation with the similarity of the target user according to the r-th targeted short video advertisement.
As a preferred aspect of the present invention, the present invention provides an optimization evaluation system according to the short video advertisement placement optimization evaluation method, including:
the hierarchy determining unit is used for acquiring a social hierarchy network representing the hierarchy social relationship of the target user according to the social communication data, clustering and quantizing the social hierarchy network into a plurality of hierarchy topologies, and marking the plurality of hierarchy topologies as incremental release hierarchies with release priorities;
the advertisement matching unit is used for carrying out similarity matching on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and taking the short video advertisements to be launched with high similarity matching degree as the quasi-launched short video advertisements which are launched to the increment launching level of the target users in increments according to the launching priority;
and the delivery volume determining unit is used for counting the advertisement conversion rate of each delivery level in the incremental delivery levels according to the delivery priority, constructing the delivery incremental relevance between the delivery levels with the high delivery priority and the low delivery priority based on the advertisement conversion rate of the delivery level with the high delivery priority, and then adjusting the delivery volume of the delivery level with the low delivery priority based on the relevance until the delivery of the quasi-delivery short video advertisement in all the delivery levels is completed according to the delivery volume.
Compared with the prior art, the invention has the following beneficial effects:
according to the method, a social hierarchy network representing the hierarchy social relationship of a target user and an incremental launching hierarchy of launching priorities are obtained according to social communication data, launching incremental relevance between the launching hierarchies with high launching priorities and low launching priorities is constructed based on the advertisement conversion rate of the launching hierarchy with high launching priorities, and then the launching amount of the launching hierarchy with low launching priorities is adjusted based on the relevance, so that the incremental adjustment of the short-video advertisement launching amount is realized to achieve the optimization of advertisement conversion rate and launching expenditure.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It should be apparent that the drawings in the following description are merely exemplary, and that other embodiments can be derived from the drawings provided by those of ordinary skill in the art without inventive effort.
Fig. 1 is a flowchart of a short video advertisement delivery optimization evaluation method according to an embodiment of the present invention;
fig. 2 is a block diagram of an evaluation system according to an embodiment of the present invention.
The reference numerals in the drawings denote the following, respectively:
1-a hierarchy determination unit; 2-an advertisement matching unit; and 3, a putting amount determining unit.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, the present invention provides a short video advertisement placement optimization evaluation method, which includes the following steps:
step S1, acquiring a social network representing the hierarchical social relationship of the target user according to the social communication data, clustering and quantizing the social network into a plurality of hierarchical topologies, and marking the plurality of hierarchical topologies as incremental release levels with release priorities;
the social network of the target user represents the extraction of the social circle of the target user, and because the social circle of the target user is usually formed by person objects with similar interest characteristics to a certain extent, the product in which the target user is interested is also the product in which the person in the social network of the target user is interested to a large extent, and the mapping to the real scene is as follows: the interest characteristics of the target user are given to friends of the target user, products (namely, the commodities interested by the target user) corresponding to the interest characteristics of the target user are recommended to friends of the target user when commodity recommendation is carried out, the generalization force of product recommendation is improved, namely, the generalization force of advertisement putting is optimized, the volume of transaction is promoted, the interest degree is quantized into putting levels, and putting priorities are given to the putting levels, wherein the higher the interest degree of the putting priority is, the earlier the putting is, the lower the interest degree of the putting priority is, the later the putting is, the specific steps are as follows:
the method for acquiring the social network representing the target user hierarchical social relationship according to the social communication data comprises the following steps:
setting a communication data volume threshold value representing the normalized social relationship, extracting a communication object of which the social communication data volume with the target user exceeds the communication data volume threshold value from the social communication data, and constructing the communication object as a 1 st level topological node representing the 1 st level social relationship of the target user;
extracting communication objects with the social communication data quantity of the topological nodes of the levels 1 to 5 exceeding a communication data quantity threshold value from the social communication data in sequence, and respectively constructing the communication objects as topological nodes of the levels 2 to 6 representing the social relations of the levels 2 to 6 of the target user;
and quantizing the target user into 0 th-level topological nodes, sequentially connecting the 0 th-6 th-level topological nodes to obtain a social hierarchical network, and taking the connection relation of the topological nodes as the edge relation of the social hierarchical network.
Because the social relationship can be fully covered by social personnel under the condition of 6 layers, 6 layers are selected when the social hierarchy network is constructed, and the social relationship can be increased or decreased when the social hierarchy network is used, so that the practical standard is met.
Clustering and quantizing the social hierarchical network into a plurality of hierarchical topologies, wherein the hierarchical topologies comprise:
clustering operation is carried out on the social hierarchical network based on modularity to obtain a plurality of topological communities, the topological communities where target users are located in the topological communities are used as the level 1 topology, and the rest topological communities are arranged in a reverse order according to the degree of a margin coefficient of the level 1 topology to be used as the level 1 topology
The method comprises the steps of (1) level topology, wherein m is represented as the total number of topological communities, and j is a metering constant;
preferably, the second step is sequentially calculated based on the amount of the side relation coefficient
The social affinity of the hierarchical topology to the level 1 topology includes:
in turn will be
The edge relation data quantity of the level topology and the level 1 topology is normalized to obtain the level one
The social intimacy between the hierarchical topology and the 1 st hierarchical topology is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
characterized by a social affinity of the jth hierarchical topology to the level 1 topology,
the number of edge relationships represented by the jth level topology and the level 1 topology.
Marking a plurality of tier topologies as incremental drop tiers having drop priorities, comprising:
in turn will be
The releasing priority of the hierarchical topology is set to be k level, the hierarchical topology with the k level releasing priority is marked as the kth releasing level with the k level releasing priority, and all releasing levels are arranged according to the priority to form an incremental releasing level together.
Step S2, similarity matching is carried out on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and the short video advertisements to be launched with high similarity matching degree are used as the short video advertisements to be launched which are launched to the target users in an increment launching level according to the launching priority;
and performing similarity matching on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, wherein the similarity matching comprises the following steps:
quantizing the consumption attribute of the target user from a semantic form into a vector form by using an NLP tool word2vec to obtain a consumption attribute vector, and quantizing the product attribute of the short video advertisement to be launched from the semantic form into the vector form by using the NLP tool word2vec to obtain a product attribute vector;
establishing a similarity matching formula of the consumption attribute of the target user and the product attribute of the short video advertisement to be launched based on the Euclidean distance of the consumption attribute vector and the product attribute vector, wherein the similarity matching formula is as follows:
in the formula (I), the compound is shown in the specification,
characterized by the similarity of the product attribute of the r-th short video advertisement to be delivered and the consumption attribute of the target user,
、
respectively representing the product attribute vector and the consumption attribute vector, wherein r is a metering constant;
setting a similarity matching threshold, and comparing the similarity of the product attribute of the short video advertisement to be delivered and the consumption attribute of the target user with the similarity matching threshold in sequence, wherein,
if the similarity is similar
If the similarity matching threshold is exceeded, the similarity is calculated
The corresponding r-th short video advertisement to be delivered is taken as a standard for delivering the short video advertisement;
if the similarity is similar
If the similarity matching threshold is not exceeded, the similarity is compared
And the corresponding r-th short video advertisement to be delivered is not taken as the short video advertisement to be delivered.
And performing accurate advertisement putting matching according to the similarity, wherein the higher the similarity is, the more the product attribute of the short video advertisement to be put conforms to the consumption attribute of the target user, the higher the product conversion possibility is, the more the product is suitable for being pushed to the target user, and the lower the similarity is, the less the product attribute of the short video advertisement to be put conforms to the consumption attribute of the target user, the lower the product conversion possibility is, the less the meaning of the product pushed to the target user is, and the watching experience of the target user is reduced.
Step S3, counting the advertisement conversion rate of each delivery level in the incremental delivery levels according to the delivery priority, constructing the delivery incremental relevance between the delivery levels with high delivery priority and low delivery priority based on the advertisement conversion rate of the delivery level with high delivery priority, and adjusting the delivery amount of the delivery level with low delivery priority based on the relevance until the delivery of the quasi-delivery short video advertisements in all the delivery levels is completed according to the delivery amount, so as to realize the incremental adjustment of the delivery amount of the short video advertisements to achieve the optimization of the advertisement conversion rate and the delivery expenditure.
Counting the advertisement conversion rate of each delivery layer in the incremental delivery layers according to the delivery priority, wherein the counting comprises the following steps:
the method comprises the following steps of counting the ratio of the number of users purchasing products in the r-th quasi-delivery short video advertisement in the k-th delivery layer of the incremental delivery layer to all the users in the k-th delivery layer in sequence to serve as the advertisement conversion rate of the r-th quasi-delivery short video advertisement in the k-th delivery layer, wherein the calculation formula of the advertisement conversion rate is as follows:
in the formula (I), the compound is shown in the specification,
characterized by the ad conversion rate of the r-th short video advertisement to be delivered in the k-th delivery level,
characterized by the number of users in the kth placement level who purchased the product in the r-th quasi-placement short video advertisement,
the characteristics are all the users in the k-th release level.
Constructing a delivery increment relevance between delivery levels with high delivery priority and low delivery priority based on the advertisement conversion rate of the delivery level with high delivery priority, comprising:
obtaining the advertisement conversion rate of the delivery layer with high delivery priority, setting an attenuation coefficient for the advertisement conversion rate of the delivery layer with high delivery priority, attenuating the advertisement conversion rate based on the attenuation coefficient to obtain a correlation coefficient representing the correlation of delivery increment, wherein,
the attenuation coefficient is obtained by constructing social affinity corresponding to the release level so as to realize attenuation according to the social affinity and improve the attenuation scientificity, and the calculation formula of the attenuation coefficient is as follows:
in the formula (I), the compound is shown in the specification,
characterized by attenuation coefficients from the k-th level of delivery to the k + 1-th level of delivery,
the social affinity is characterized by the k +1 th release level and the 1 st release level;
the correlation coefficient is calculated by the formula:
in the formula (I), the compound is shown in the specification,
the characteristic is that the r-th short video advertisement to be delivered is delivered with the association coefficient of the increment in the k-th delivery level and the k + 1-th delivery level;
the release increment is characterized by the increasing rate of the release amount of the k +1 level release level on the basis of the release amount of the k level release level.
Adjusting a placement amount for a placement tier having a low placement priority based on relevance, comprising:
according to the correlation coefficient, obtain
The amount of the level of delivery, the second
The calculation formula of the putting amount of the level putting level is as follows:
in the formula (I), the compound is shown in the specification,
、
respectively representing the putting quantities of the r-th quasi-put short video advertisement in the k +1 th and k-th putting levels;
setting the throwing amount of the 1 st level throwing level according to the order
And calculating the putting amount of the r-th quasi-delivery short video advertisement in the 1 st to m-th delivery levels by using a calculation formula of the putting amount of the level delivery levels.
The lower the level of placement priority, the lower the level of interest to those in the placement hierarchy for placing short video advertisements, the advertisement conversion rate will be lower, so directly delivering the placement volume in the placement level with high placement priority to the placement level with low placement priority will cause the redundancy of placement volume, the placement volume of a drop tier with a low drop priority is thus based on the decay of the placement volume of a drop tier with a high drop priority, and the construction of the attenuation coefficient is established on the advertisement conversion rate of the delivery level with high delivery priority, the delivery amount can be reduced under the condition of ensuring that the advertisement conversion rate is not influenced, the method reduces the cost of the delivery expenditure, ensures the advertisement conversion rate, and realizes the increment adjustment of the short video advertisement delivery amount to achieve the optimization of the advertisement conversion rate and the delivery expenditure.
The putting quantity of the level 1 putting level is in a direct proportion relation with the similarity of the target user according to the r-th quasi-putting short video advertisement.
As shown in fig. 2, based on the short video advertisement delivery optimization evaluation method, the present invention provides an optimization evaluation system, which includes:
the hierarchy determining unit 1 is used for acquiring a social hierarchy network representing a hierarchy social relationship of a target user according to social communication data, clustering and quantizing the social hierarchy network into a plurality of hierarchy topologies, and marking the plurality of hierarchy topologies as incremental release hierarchies with release priorities;
the advertisement matching unit 2 is used for carrying out similarity matching on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and taking the short video advertisements to be launched with high similarity matching degree as the quasi-launched short video advertisements which are launched to the increment launching level of the target users in increments according to the launching priority;
and the putting quantity determining unit 3 is used for counting the advertisement conversion rate of each putting layer in the incremental putting layers according to the putting priority, constructing the putting increment relevance between the putting layers with the high putting priority and the putting layers with the low putting priority based on the advertisement conversion rate of the putting layer with the high putting priority, and then adjusting the putting quantity of the putting layer with the low putting priority based on the relevance until the putting of the quasi-put short video advertisement in all the putting layers is finished according to the putting quantity.
According to the method, a social hierarchy network representing the target user hierarchy social relationship and an incremental delivery hierarchy of delivery priorities are obtained according to social communication data, delivery incremental relevance between the delivery hierarchies with high delivery priorities and low delivery priorities is constructed based on the advertisement conversion rate of the delivery hierarchy with high delivery priorities, and the delivery amount of the delivery hierarchy with low delivery priorities is adjusted based on the relevance, so that the incremental adjustment of the short-video advertisement delivery amount is realized to achieve the optimization of advertisement conversion rate and delivery expenditure.
The above embodiments are only exemplary embodiments of the present application, and are not intended to limit the present application, and the protection scope of the present application is defined by the claims. Various modifications and equivalents may be made by those skilled in the art within the spirit and scope of the present application and such modifications and equivalents should also be considered to be within the scope of the present application.
Claims (8)
1. A short video advertisement putting optimization evaluation method is characterized by comprising the following steps:
step S1, acquiring a social network representing the hierarchical social relationship of the target user according to the social communication data, clustering and quantizing the social network into a plurality of hierarchical topologies, and marking the plurality of hierarchical topologies as incremental release levels with release priorities;
step S2, similarity matching is carried out on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and the short video advertisements to be launched with high similarity matching degree are used as quasi-launched short video advertisements which are launched to the increment launching level of the target users in an increment launching mode according to launching priorities;
step S3, counting the advertisement conversion rate of each delivery level in the incremental delivery levels according to the delivery priority, constructing the delivery incremental relevance between the delivery levels with high delivery priority and low delivery priority based on the advertisement conversion rate of the delivery level with high delivery priority, and adjusting the delivery amount of the delivery level with low delivery priority based on the relevance until the delivery of the quasi-delivery short video advertisement in all the delivery levels is completed according to the delivery amount, so as to realize the incremental adjustment of the delivery amount of the short video advertisement to achieve the optimization of the advertisement conversion rate and the delivery expenditure;
obtaining the advertisement conversion rate of the delivery layer with high delivery priority, setting an attenuation coefficient for the advertisement conversion rate of the delivery layer with high delivery priority, attenuating the advertisement conversion rate based on the attenuation coefficient to obtain a correlation coefficient representing the relevance of the delivery increment, wherein,
the attenuation coefficient is obtained by constructing social affinity corresponding to the release level so as to realize attenuation according to the social affinity and improve attenuation scientificity, and the calculation formula of the attenuation coefficient is as follows:
in the formula, S k k,+1Characterization ofAttenuation coefficient from the k-th level of release to the k + 1-th level of release, J k+1The social affinity is characterized by the social affinity of the k +1 th release level and the 1 st release level;
the calculation formula of the correlation coefficient is as follows:
Y k k r,+1,=Q k r,*S k k,+1；
in the formula, Y k k r,+1,The characteristic is that the r-th short video advertisement to be delivered is delivered with the association coefficient of the increment in the k-th delivery level and the k + 1-th delivery level;
the release increment is characterized by the increasing rate of the release amount of the k + 1-level release level on the basis of the release amount of the k-level release level;
obtaining the putting quantity of a k (k belongs to [1, m ]) level putting hierarchy according to the correlation coefficient, wherein the calculation formula of the putting quantity of the k (k belongs to [1, m ]) level putting hierarchy is as follows:
Z k r+1,=Z k r,*Y k k r,+1,；
in the formula, Z k r+1,、Z k r,Respectively representing the putting quantities of the r-th quasi-put short video advertisement in the k +1 th and k-th putting levels;
and setting the putting quantity of the 1 st level putting level, and calculating the putting quantity of the r < th > placement-ready short video advertisement in the 1 st level to the m < th > level putting level according to the calculation formula of the putting quantity of the k (k belongs to [1, m ]) level putting level in sequence.
2. The method of claim 1, wherein the method comprises: the method for acquiring the social network representing the target user hierarchical social relationship according to the social communication data comprises the following steps:
setting a communication data volume threshold value representing a normalized social relationship, extracting a communication object of which the social communication data volume with a target user exceeds the communication data volume threshold value from the social communication data, and constructing the communication object as a level 1 topological node representing the level 1 social relationship of the target user;
extracting communication objects with the social communication data quantity of the 1 st to 5 th level topology nodes exceeding a communication data quantity threshold value from the social communication data in sequence, and respectively constructing the communication objects as 2 nd to 6 th level topology nodes representing the 2 nd to 6 th level social relations of the target user;
quantizing the target user into 0 th-level topological nodes, sequentially connecting 0 th-6 th-level topological nodes to obtain the social network, and taking the connection relation of the topological nodes as the edge relation of the social network.
3. The method of claim 2, wherein the method comprises: clustering quantization is performed on the social hierarchical network into a plurality of hierarchical topologies, and the clustering quantization comprises the following steps:
clustering operation is carried out on the social hierarchical network based on modularity to obtain a plurality of topological communities, the topological communities where target users are located in the topological communities are used as a level 1 topology, the rest topological communities are arranged in a reverse order according to the relationship coefficient quantity of the level 1 topology to be used as a level j (j belongs to [2, m ]) topology, m is represented as the total number of the topological communities, and j is a metering constant;
sequentially calculating the social affinity of the jth (j belongs to [2, m ]) level topology and the 1 st level topology based on the relationship coefficient quantity, wherein the social affinity comprises the following steps:
sequentially carrying out normalization processing on the edge relation data quantity of the jth (j belongs to [2, m ]) level topology and the 1 st level topology to obtain the social affinity of the jth (j belongs to [2, m ]) level topology and the 1 st level topology, wherein the social affinity has the calculation formula:
in the formula, J j Characterized by a social affinity of the jth hierarchical topology to the level 1 topology,M j the number of edge relationships represented by the jth level topology and the level 1 topology.
4. The method of claim 3, wherein the method comprises: the marking of the multiple tier topology as an incremental drop tier with drop priority comprises:
and sequentially setting the launching priority of the kth (k belongs to [1, m ]) hierarchical topology as k level, marking the hierarchical topology with the k level launching priority as the kth launching level with the k level launching priority, and arranging all launching levels according to the priority to form an incremental launching level.
5. The method of claim 4, wherein the method comprises: the similarity matching of the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence comprises the following steps:
quantizing the consumption attribute of the target user from a semantic form into a vector form by using an NLP tool word2vec to obtain a consumption attribute vector, and quantizing the product attribute of the short video advertisement to be delivered from the semantic form into the vector form by using the NLP tool word2vec to obtain a product attribute vector;
establishing a similarity matching formula of the consumption attribute of the target user and the product attribute of the short video advertisement to be launched based on the Euclidean distance of the consumption attribute vector and the product attribute vector, wherein the similarity matching formula is as follows:
in the formula I r Characterized by the similarity of the product attribute of the r-th short video advertisement to be delivered and the consumption attribute of the target user,X r 、xrespectively representing the product attribute vector and the consumption attribute vector, wherein r is a metering constant;
setting a similarity matching threshold, and comparing the similarity of the product attribute of the short video advertisement to be launched and the consumption attribute of the target user with the similarity matching threshold in sequence, wherein,
if the similarity I r Exceeding the similarity matchMatching the threshold value, the similarity I r The corresponding r-th short video advertisement to be delivered is taken as a standard for delivering the short video advertisement;
if the similarity I r If the similarity matching threshold is not exceeded, the similarity I is calculated r And the corresponding r-th short video advertisement to be delivered is not taken as the short video advertisement to be delivered.
6. The method of claim 5, wherein the method comprises: counting the advertisement conversion rate of each delivery layer in the incremental delivery layers according to the delivery priority, wherein the counting comprises the following steps:
counting the ratio of the number of users who purchase products in the r-th quasi-delivery short video advertisement in the k-th delivery layer of the incremental delivery layer to the number of all users in the k-th delivery layer in turn as the advertisement conversion rate of the r-th quasi-delivery short video advertisement in the k-th delivery layer, wherein the calculation formula of the advertisement conversion rate is as follows:
in the formula, Q k r,Characterized by the ad conversion rate of the r-th short video advertisement to be delivered in the k-th delivery level,N k r,characterized by the number of users in the kth placement level who purchased the product in the r-th quasi-placement short video advertisement,N k the characteristics are all the users in the k-th release level.
7. The method according to claim 6, wherein the placement amount of the level 1 placement level is proportional to the similarity between the mth placement-ready short video advertisement and the target user.
8. An optimization evaluation system of the short video advertisement placement optimization evaluation method according to any one of claims 1 to 7, comprising:
the system comprises a level determining unit (1) and a releasing unit, wherein the level determining unit is used for acquiring a social level network representing the level social relationship of a target user according to social communication data, clustering and quantizing the social level network into a plurality of level topologies, and marking the plurality of level topologies as incremental releasing levels with releasing priorities;
the advertisement matching unit (2) is used for carrying out similarity matching on the consumption attributes of the target users and the product attributes of the short video advertisements to be launched in sequence, and taking the short video advertisements to be launched with high similarity matching degree as the quasi-launched short video advertisements which are launched in increments according to launching priorities to the increment launching level of the target users;
the delivery volume determining unit (3) is used for counting the advertisement conversion rate of each delivery level in the incremental delivery levels according to the delivery priority, constructing the delivery incremental relevance between the delivery levels with the high delivery priority and the low delivery priority based on the advertisement conversion rate of the delivery level with the high delivery priority, and then adjusting the delivery volume of the delivery level with the low delivery priority based on the relevance until the delivery of the quasi-delivery short video advertisement in all the delivery levels is completed according to the delivery volume;
obtaining the advertisement conversion rate of the delivery layer with high delivery priority, setting an attenuation coefficient for the advertisement conversion rate of the delivery layer with high delivery priority, attenuating the advertisement conversion rate based on the attenuation coefficient to obtain a correlation coefficient representing the relevance of the delivery increment, wherein,
the attenuation coefficient is obtained by constructing social affinity corresponding to the release level so as to realize attenuation according to the social affinity and improve attenuation scientificity, and the calculation formula of the attenuation coefficient is as follows:
in the formula, S k k,+1Characterized by attenuation coefficients from the k-th level of delivery to the k + 1-th level of delivery, J k+1The social affinity is characterized by the social affinity of the k +1 th release level and the 1 st release level;
the calculation formula of the correlation coefficient is as follows:
Y k k r,+1,=Q k r,*S k k,+1；
in the formula, Y k k r,+1,The characteristic is that the r-th short video advertisement to be delivered is delivered with the association coefficient of the increment in the k-th delivery level and the k + 1-th delivery level;
the release increment is characterized by the increasing rate of the release amount of the k + 1-level release level on the basis of the release amount of the k-level release level;
obtaining the putting quantity of a k (k belongs to [1, m ]) level putting hierarchy according to the correlation coefficient, wherein the calculation formula of the putting quantity of the k (k belongs to [1, m ]) level putting hierarchy is as follows:
Z k r+1,=Z k r,*Y k k r,+1,；
in the formula, Z k r+1,、Z k r,Respectively representing the putting quantities of the r-th quasi-put short video advertisement in the k +1 th and k-th putting levels;
and setting the putting quantity of the 1 st level putting level, and calculating the putting quantity of the r < th > placement-ready short video advertisement in the 1 st level to the m < th > level putting level according to the calculation formula of the putting quantity of the k (k belongs to [1, m ]) level putting level in sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210203897.2A CN114266606B (en) | 2022-03-03 | 2022-03-03 | Short video advertisement putting optimization evaluation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210203897.2A CN114266606B (en) | 2022-03-03 | 2022-03-03 | Short video advertisement putting optimization evaluation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114266606A CN114266606A (en) | 2022-04-01 |
| CN114266606B true CN114266606B (en) | 2022-05-17 |
Family
ID=80834014
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210203897.2A Active CN114266606B (en) | 2022-03-03 | 2022-03-03 | Short video advertisement putting optimization evaluation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114266606B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109711885A (en) * | 2018-12-27 | 2019-05-03 | 上海旺翔文化传媒股份有限公司 | Motivate video ads intelligence put-on method |
| CN111598633A (en) * | 2020-07-24 | 2020-08-28 | 北京淇瑀信息科技有限公司 | Online advertisement putting method and device based on incremental learning and electronic equipment |
| CN112435067A (en) * | 2020-11-30 | 2021-03-02 | 翼果（深圳）科技有限公司 | Intelligent advertisement putting method and system for cross-e-commerce platform and social platform |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110264522A1 (en) * | 2010-04-26 | 2011-10-27 | Webjuice, LLC | Direct targeting of advertisements to social connections in a social network environment |
| US8887197B2 (en) * | 2012-11-29 | 2014-11-11 | At&T Intellectual Property I, Lp | Method and apparatus for managing advertisements using social media data |
| US20160180375A1 (en) * | 2014-12-22 | 2016-06-23 | Lochlan H. Rose | System And Method To Estimate The Incrementality Delivered By Online Campaigns Based On Measuring Impression-Level Digital Display Ad Viewability |
-
2022
- 2022-03-03 CN CN202210203897.2A patent/CN114266606B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109711885A (en) * | 2018-12-27 | 2019-05-03 | 上海旺翔文化传媒股份有限公司 | Motivate video ads intelligence put-on method |
| CN111598633A (en) * | 2020-07-24 | 2020-08-28 | 北京淇瑀信息科技有限公司 | Online advertisement putting method and device based on incremental learning and electronic equipment |
| CN112435067A (en) * | 2020-11-30 | 2021-03-02 | 翼果（深圳）科技有限公司 | Intelligent advertisement putting method and system for cross-e-commerce platform and social platform |
Non-Patent Citations (2)
| Title |
|---|
| Advertising in Popular Instant Messengers;D. Lugovoy等;《2018 IEEE Communication Strategies in Digital Society Workshop (ComSDS)》;20180507;第39-41页 * |
| 基于增量学习的精准广告投放系统研究;李娜;《硕士电子期刊》;20110415;第1-49页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114266606A (en) | 2022-04-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9727927B2 (en) | Prediction of user response to invitations in a social networking system based on keywords in the user's profile | |
| US20190279230A1 (en) | Online Content Delivery Based on Information from Social Networks | |
| CN104090967B (en) | Application program recommends method and recommendation apparatus | |
| CN104915392B (en) | A kind of microblogging forwarding behavior prediction method and device | |
| CN106372959A (en) | Internet-based user access behavior digital marketing system and method | |
| CN107222566A (en) | Information-pushing method, device and server | |
| CN106504098A (en) | A kind of capitalized method of the financial investment system based on big data technology | |
| CN106776873A (en) | A kind of recommendation results generation method and device | |
| CN108777701A (en) | A kind of method and device of determining receiver | |
| CN109697627A (en) | System and method for using deep layer nerve language model to bid automatically | |
| US20200098015A1 (en) | Methods for determining targeting parameters and bids for online ad distribution | |
| CN106886559A (en) | The collaborative filtering method of good friend's feature and similar users feature is incorporated simultaneously | |
| CN107545444A (en) | A kind of card data recommendation method and device | |
| CN110191171A (en) | Weather information transmission method | |
| CN114565407A (en) | Advertisement delivery data analysis method and system | |
| CN114266606B (en) | Short video advertisement putting optimization evaluation method and system | |
| CN107507023B (en) | Information delivery method and device | |
| CN106204163B (en) | Method and device for determining user attribute characteristics | |
| CN105208033B (en) | A kind of colony's auxiliary based on intelligent terminal scene recommends method and system | |
| CN107463853A (en) | The method and system of audient's label analysis | |
| CN114565408B (en) | Bidding prediction method and system for advertisement putting | |
| CN113362095A (en) | Information delivery method and device | |
| CN114155067B (en) | User portrait establishing method and system | |
| CN116226537A (en) | Layout and display method, device, equipment and medium of page modules in page | |
| CN110070392A (en) | Customer churn method for early warning and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |