35 results about "Computational advertising" patented technology

A mobile advertising system includes a banner, a first banner anchor and a second banner anchor coupled to a vehicle. The banner is capable of being deployed by being connected at opposing ends to and suspended between the first and second banner anchors. The banner has first and second surfaces on which indicia are imprinted. Both the first and second banner surfaces are visible when the banner is deployed. An invoicing and tracking system may be included that computes an advertising charge based at least in part on a time period during which the banner is deployed and, optionally, a location of the vehicle.

The invention discloses a method and device for automatically managing advertisement putting and electronic equipment. The method comprises the steps: obtaining the feature data of a plurality of historical advertisement putting and the corresponding actual effect data, enabling the feature data to comprise putting channel data and putting mode data, and enabling the feature data to serve as training data; establishing a machine self-learning model for calculating an advertisement putting effect, and training the machine self-learning model by using the training data; predicting the obtained actual effect of the feature data of the to-be-delivered advertisement, and generating suggestion data; and modifying the feature data of the advertisement to be released according to the suggestion data. Advertisement putting channels and putting modes are obtained from historical advertisement putting to serve as feature data, the actual effect of historical advertisement putting is quantified into effect data, training of a machine self-learning model is carried out, after the feature data of the to-be-put advertisement is input into the model, prediction of the effect data can be achieved,and according to the quality of a prediction result, the advertisement putting effect is improved. Therefore, the characteristic data of the advertisement can be adjusted.

The invention discloses a marketing activity prediction method based on knowledge distillation. The method comprises the steps of data preprocessing, data set division and network training framework formation of a teacher model, data set division and network training framework formation of a student model, prediction model establishment, marketing activity prediction and the like. The method includes following steps: firstly, constructing a relatively complex teacher model Net-T with a residual neural network as a core, and then establishing a student model Net-S formed by a simple neural network; weighting a soft label obtained by training a teacher model Net-T at a high temperature and a hard label obtained by training a student model Net-S at the same temperature to obtain a total loss function of knowledge distillation; by taking the total loss function as a target function of a student model Net-S during actual deployment, training to obtain a final neural network model, and performing prediction, wherein the result shows that the hybrid model effectively expands the application of deep learning to advertisement calculation and recommendation system algorithms, so that the method significantly improves the accuracy of user click prediction.

The invention discloses an outdoor advertisement recommendation method based on large-scale user portraits, and belongs to the technical field of computational advertising. According to the outdoor advertisement recommendation method, an advertisement scheduling module is connected with a user side, an advertisement management module and a real-time accurate crowd analysis module. And the real-time accurate crowd analysis module performs parameter interaction with the advertisement retrieval module, the user portrait module and the user behavior module. The user data mining module is connectedwith the advertisement management module, the user behavior module and the user portrait module, and the advertisement management module is connected with the advertisement retrieval module. According to the outdoor advertisement recommendation method, when the user logs in the platform, the user interest and the historical behavior record are inquired according to the user portrait to generate the personalized advertisement list, and finally the advertisement is pushed to the outdoor advertisement terminal. The method is good in autonomous learning ability, can effectively improve the touchrate of outdoor intelligent advertisements, and is suitable for outdoor advertisement recommendation under the background of the Internet of Things and big data.

The invention relates to an SIFT-based advertisement material and on-site picture automatic matching evaluation method, a storage medium and computer equipment, and the method comprises the steps: carrying out the feature point extraction of an advertisement material picture and an on-site picture, carrying out the quick matching of feature points, and removing the feature points of which the distance is greater than a preset distance range; if the number of the feature point groups meeting the distance range reaches four or more, calculating a homography matrix of the advertisement material picture feature points mapped to the field picture feature points; mapping the vertex coordinates of the advertisement material picture to the on-site picture to obtain a polygon A formed by the mappedvertex set, forming a polygon B by the vertex set of the field picture, and calculating the intersection of the two polygons A and B on the on-site picture to form a polygon S; and judging whether the polygon S is a convex quadrangle, and if so, calculating an automatic matching evaluation score of the shooting quality of the on-site picture through a preset evaluation model. According to the invention, the cost of manual auditing is effectively reduced, and the problem of inconsistent auditing standards of different personnel is solved.

The invention provides an advertisement system for a meta universe, which comprises a plurality of advertisement positions arranged in the meta universe, advertisement entities are inserted into the advertisement positions, when the advertisement entities effectively collide with the view field of an advertisement target user, the system starts timing, and when the timing time reaches a preset duration, the advertisement target user is started to enter the advertisement position. The system begins to calculate advertising revenue involving the advertising entity. The advertisement entities are of a three-dimensional structure, the advertisement entities are arranged into a plurality of categories according to an arrangement mode, the advertisement positions are correspondingly arranged into a plurality of categories, the system compares whether the advertisement entities are matched with the categories of the advertisement positions or not, and if yes, the advertisement entities can be inserted into the advertisement positions. According to the method, excessive release of the advertisement is limited through the preset advertisement position, the experience of the user in the virtual world can be ensured, the advertisement revenue is calculated through a scheme of judging that the view range of the advertisement target user effectively collides with the advertisement entity and timing, and the benefit between an advertisement publisher and an advertisement customer can be accurately balanced.

The invention discloses an advertisement putting method. The method comprises the following steps: acquiring terminal information uploaded by a traffic owner; receiving an advertisement putting instruction triggered by an advertiser according to the terminal information, and calculating advertisement putting cost according to the advertisement putting instruction; if it is detected that the advertisement putting cost is paid, putting advertisement information needing to be put by an advertiser to the target terminal for playing through a set top box connected with the target terminal in response to the advertisement putting instruction, wherein the target terminal corresponds to the advertisement putting instruction, and is used for playing the advertisement information released by the advertiser. The invention further discloses an advertisement putting system and a computer readable storage medium. The terminal information is uploaded autonomously through the flow owner, the advertiser triggers the advertisement putting instruction autonomously, so that the advertisement putting is more flexible, the set top box is controlled to enable the target terminal to play the put advertisement information, the advertisement can be put anytime and anywhere, and the advertisement putting effect is improved.

The invention discloses a marketing prediction method combining inner/outer product feature interaction and a Bayesian neural network. The method comprises a data preprocessing step, a data set division step, a model establishment step and a marketing activity click prediction step. According to the method, in the building process of a prediction model, the prediction uncertainty is introduced into the Bayesian neural network by effectively utilizing Bayesian inference, so that the Bayesian neural network model has higher robustness. And by adopting an inner/outer product combination method, the features are crossed to extract high-dimensional recessive features. Therefore, the application of deep learning to advertisement calculation and recommendation system algorithm problems can be effectively expanded, and the accuracy of user click behavior prediction is remarkably improved.

A decentralized and trust-minimizing computer architecture for computing rewards for users of an advertising system includes cryptographic black box accumulators (BBA), which is a cryptographic counter that only the issuer can update. An attention application requests initialization of a BBA from a guardian and subsequently requests updates to the BBA to track interactions between a user of the attention application and ads on the attention application. The guardian signs updates to the BBA to reach agreement on the state of ad interactions. The attention application may randomize the BBA and submit requests via an anonymous channel such that no participant can link two encounters with the BBA to each other or link the BBA to a specific attention application, thus improving user privacy. Reward redemption requests can be made based on a known policy and committed to a public blockchain for verification by observers that the protocol is operating correctly.

The invention relates to an advertisement media contact person number detection device and an advertisement media contact person number detection method. The advertisement media contact person numberdetection device comprises a visual identification range calculation unit, a block extraction unit, a block elimination unit, a range designation unit and a contact person number calculation unit, wherein floating population data is acquired and the advertisement media contact person number is calculated through the contact person number calculation unit. According to the invention, actual trafficsurvey is not needed, and the number of media contact persons in any period can be accurately and easily detected at any time.

A computer-implemented method comprises receiving, from an ad exchange server via a data communications network, a message comprising a bid request which includes site information and user informationrelating to an available ad slot. For each offer in ranked list generated, an offer-level estimate of probability of user interaction with the offer is computed. For a combination of offers includedin the ranked list, an ad-level bid price is computed, based on at least the computed offer-level estimates of probability of user interaction, corresponding offer-level interaction revenues, and an aggressiveness parameter. A bid response includes a bid-priced ad which comprises the combination of offers and the ad-level bid price. Machine learning models for predicting behaviour of online usersare able to automatically determine estimates of probability of user interaction with online content elements based upon aggregated behaviour of prior users in similar contexts.

An apparatus for detecting the number of persons contacted with an advertising medium, includes a visual-recognition range calculation unit for calculating a visual recognition ground range of the advertising medium based on a position of the advertising medium, a height above the ground of the advertising medium and a direction of an advertising surface of the advertising medium, a section extraction unit for extracting a plurality of sections in an area allowed persons to pass through, a section exclusion unit for excluding at least one section that is impossible to visually recognize the advertising medium from the plurality of sections extracted, an area specifying unit for specifying an area that contains the remaining sections except the at least one section excluded, and a contacted number calculation unit for calculating the number of persons contacted with the advertising medium by obtaining fluid population data of persons moving toward the advertising.

The invention discloses an advertisement putting method and system, and the method comprises the following steps: judging whether the data of a putting list is updated or not in a time slice; if yes,acquiring delivery order data through a search server; analyzing the delivery order data, and obtaining statistical information in the delivery order data, the statistical information comprising the balance of the advertiser, the pre-calculated amount of the day, the expenditure amount of the day, the expenditure amount of the last time slice, the delivery speed of the last time slice and the remaining time slice; calculating the putting speed of the next time slice according to the calculation function and the putting list data; and writing the putting speed of the next time slice into the putting list data. The advertisement putting speed can be automatically calculated, the advertisement putting speed can be adjusted in time, the pre-calculated amount is matched with the putting speed,and the optimal advertisement putting effect is achieved.

The invention provides a method and device for determining the advertisement conversion rate, a storage medium and electronic equipment applied to an advertisement putting platform, and the method comprises the steps of receiving a trigger request for calculating the advertisement conversion rate; in response to the trigger request, obtaining first data, second data and a flow growth rate; and determining an advertisement conversion rate according to the first data, the second data and the flow growth rate. It can be seen that after a trigger request for calculating the advertisement conversion rate is received, the number of clicks used for clicking a preset event, the number of clicks of an advertisement in the advertisement putting platform and the flow increasing rate can be obtained according to the trigger request so that when account data information under a third-party platform cannot be obtained, the advertisement conversion rate can still be calculated.

The invention relates to a method and device for regulating and controlling market putting cost and computer equipment. The method comprises the following steps: collecting channel cost, order profit data and profit calculation dimension data; then, calculating the profit rate of the current channel under the current rule according to a set analysis model rule, the channel cost, the order profit data and the profit calculation dimension data; and finally, regulating the market delivery amount according to profit expectation and the profit rate. According to the scheme, the effect of dynamically calculating the advertisement putting in real time can be achieved, the advertisement putting amount is automatically set according to the calculation result, repeated operation of manual analysis and manual advertisement putting setting is omitted, the labor cost is reduced, and stable increase of profits is guaranteed by more timely and accurately controlling the putting cost.

Generating a cannibalism score for a paid ad in a search engine results page (SERP) by gathering keywords relevant to an advertiser, defining rules that compute a cannibalism score for the ad in relation to a corresponding unpaid listing, where the cannibalism score estimates the reduction in revenue to the advertiser due to the ad appearing in the same SERP as the corresponding listing, providing a keyword to a search engine, receiving a SERP from the search engine, determining the position of a first ad placed by the advertiser from among one or more ads in the SERP, determining the position of a corresponding unpaid listing from among a plurality of unpaid listings in the SERP, and applying the rules to the ad and to the corresponding unpaid listing to obtain a cannibalism score for the ad.

A click rate estimation model for performing feature interaction selection based on a three-way decision theory is used for calculating a click rate estimation task in an advertisement and recommendation system, and the method comprises the following steps of: 1) processing input classification features by utilizing One-hot coding and an Embedding technology; 2) interacting the characteristics of the original data by using a factorization machine and a derivative model thereof; 3) selecting interaction features by using a three-way decision gate proposed by combining a three-way decision theory with a binary Sigmoid function; and 4) outputting the finally predicted click rate through a logistic regression function according to the interactive characteristics selected by the three decision gates and the original characteristics of the data. Noise information brought to an original model by excessive redundant interaction features can be reduced, important feature interaction is enhanced to a certain extent, general important interaction features are reserved, and the redundant interaction features are eliminated. The performance of the factorization machine and the derivative model thereof is improved, and meanwhile the training time of the model is shortened.

The invention provides an advertisement frame value calculation device and an advertisement frame value calculation method. The value of an advertisement frame using a digital signage as an advertisement medium can be calculated more appropriately. Provided is an advertisement box value calculation device for calculating the value of an advertisement box based on a digital signage, the advertisement box value calculation device comprising: a position information acquisition unit for acquiring information on the position where the digital signage is provided; and a region information acquisition means for acquiring region-related information in a region including the position where the digital signage is provided from a regularly or irregularly updated database in which the region-related information is stored, the area-related information is information related to factors that affect the increase and decrease of the number of staying persons in each of a plurality of predetermined areas; and a value calculation unit that calculates the value of the advertisement box on the basis of the acquired region-related information.

A method for classifying ad networks to be used by an app, comprising the following steps performed by a server (10):

• • for each app in a database (12), computing a quality index of the app based on evaluating parameters related to the app using an ad network and related to other apps using the same ad network;
  • for each app using the ad network, computing an aggressivity level of the ad network based on aggressivity parameters which measure intrusivity, impact and effectivity of the ad network in the app;
  • ranking all the apps of the database (12) according to the computed quality index and aggressivity level;
  • for each ad network, computing a single classification metric based on structural parameters related to the ad network and to the apps using it;
  • ranking all the ad networks according to the computed classification metric;
  • delivering the rankings of the mobile applications and ad networks to both private users (14) and professional users (15).