广告点击率是互联网广告投放的重要依据,有效地预测广告的点击率,对于提高广告投放的效率有着至关重要的作用。在训练点击率预测模型的过程中,往往面临着广告及用户的数量巨大以及训练数据集稀疏的问题,从而导致点击率预测的准确度下降...广告点击率是互联网广告投放的重要依据,有效地预测广告的点击率,对于提高广告投放的效率有着至关重要的作用。在训练点击率预测模型的过程中,往往面临着广告及用户的数量巨大以及训练数据集稀疏的问题,从而导致点击率预测的准确度下降。针对这些问题提出了一种基于LDA(latent Dirichlet allocation,LDA)的点击率预测算法,即LDA-FMs,该算法对原有训练集进行基于主题的分割,利用分割后的子训练集分别建立不同主题下的点击率预测模型;在此基础上,利用广告属于不同主题的概率,有权重地结合每个预测模型的预测结果,进而计算广告的点击率。实验基于KDD Cup 2012-track2的真实数据集,证明了算法的可行性与有效性。展开更多
点击率预测可以提高用户对所展示互联网广告的满意度,支持广告的有效投放,是针对用户进行广告的个性化推荐的重要依据.对于没有历史点击记录的用户,仍需对其推荐广告,预测所推荐广告的点击率.针对这类用户,以贝叶斯网这一重要的概率图模...点击率预测可以提高用户对所展示互联网广告的满意度,支持广告的有效投放,是针对用户进行广告的个性化推荐的重要依据.对于没有历史点击记录的用户,仍需对其推荐广告,预测所推荐广告的点击率.针对这类用户,以贝叶斯网这一重要的概率图模型,作为不同用户之间广告搜索行为的相似性及其不确定性的表示和推理框架,通过对用户搜索广告的历史记录进行统计计算,构建反映用户间相似关系的贝叶斯网,进而基于概率推理机制,定量度量没有历史点击记录的用户与存在历史点击记录的用户之间的相似性,从而预测没有历史点击记录的用户对广告的点击率,为广告推荐提供依据.通过建立在KDD Cup 2012-Track 2的Tencent CA训练数据集上的实验,测试了方法的有效性.展开更多
In the era of intelligent economy, the click-through rate(CTR) prediction system can evaluate massive service information based on user historical information, and screen out the products that are most likely to be fa...In the era of intelligent economy, the click-through rate(CTR) prediction system can evaluate massive service information based on user historical information, and screen out the products that are most likely to be favored by users, thus realizing customized push of information and achieve the ultimate goal of improving economic benefits. Sequence modeling is one of the main research directions of CTR prediction models based on deep learning. The user's general interest hidden in the entire click history and the short-term interest hidden in the recent click behaviors have different influences on the CTR prediction results, which are highly important. In terms of capturing the user's general interest, existing models paid more attention to the relationships between item embedding vectors(point-level), while ignoring the relationships between elements in item embedding vectors(union-level). The Lambda layer-based Convolutional Sequence Embedding(LCSE) model proposed in this paper uses the Lambda layer to capture features from click history through weight distribution, and uses horizontal and vertical filters on this basis to learn the user's general preferences from union-level and point-level. In addition, we also incorporate the user's short-term preferences captured by the embedding-based convolutional model to further improve the prediction results. The AUC(Area Under Curve) values of the LCSE model on the datasets Electronic, Movie & TV and MovieLens are 0.870 7, 0.903 6 and 0.946 7, improving 0.45%, 0.36% and 0.07% over the Caser model, proving the effectiveness of our proposed model.展开更多
In the era of computational advertising,advertising effectiveness can be measured by different metrics at different stages of the sales funnel.In the upper funnel,the click-through rate(CTR,the rate of click per impre...In the era of computational advertising,advertising effectiveness can be measured by different metrics at different stages of the sales funnel.In the upper funnel,the click-through rate(CTR,the rate of click per impression)represents the attractiveness of the advertising;the conversion rate(CVR,the rate of conversion per click)in the lower funnel indicates the persuasiveness of the advertising.Achieving higher CTR and CVR may need distinct advertising strategies:improving CTR requires raising more consumers'interest in the ad,which is more beneficial to publishers;boosting CVR needs the ad to inspire more consumers’desire in the product(service),which is more profitable to advertisers.In order to study the performance of advertising texts in terms of the two dimensions and reconcile the two different goals,this paper draws on Speech act theory(SAT)in linguistics to classify advertising texts into three types(i.e.,assertive,expressive,and directive),and analyzes how advertising texts can impact consumer behaviors.We further categorize the above three styles of advertising texts into subjective type(i.e.,expressive and directive)and objective type(i.e.,assertive).Based on a field study,we find that subjective advertising with more personalization leads to a higher CTR,while objective advertising with higher consistency with the brand information results in a higher CVR.The results suggest that firms with different marketing goals should utilize different styles of advertising texts to elicit desirable consumer behaviors during different stages of the sales funnel.展开更多
文摘广告点击率是互联网广告投放的重要依据,有效地预测广告的点击率,对于提高广告投放的效率有着至关重要的作用。在训练点击率预测模型的过程中,往往面临着广告及用户的数量巨大以及训练数据集稀疏的问题,从而导致点击率预测的准确度下降。针对这些问题提出了一种基于LDA(latent Dirichlet allocation,LDA)的点击率预测算法,即LDA-FMs,该算法对原有训练集进行基于主题的分割,利用分割后的子训练集分别建立不同主题下的点击率预测模型;在此基础上,利用广告属于不同主题的概率,有权重地结合每个预测模型的预测结果,进而计算广告的点击率。实验基于KDD Cup 2012-track2的真实数据集,证明了算法的可行性与有效性。
文摘点击率预测可以提高用户对所展示互联网广告的满意度,支持广告的有效投放,是针对用户进行广告的个性化推荐的重要依据.对于没有历史点击记录的用户,仍需对其推荐广告,预测所推荐广告的点击率.针对这类用户,以贝叶斯网这一重要的概率图模型,作为不同用户之间广告搜索行为的相似性及其不确定性的表示和推理框架,通过对用户搜索广告的历史记录进行统计计算,构建反映用户间相似关系的贝叶斯网,进而基于概率推理机制,定量度量没有历史点击记录的用户与存在历史点击记录的用户之间的相似性,从而预测没有历史点击记录的用户对广告的点击率,为广告推荐提供依据.通过建立在KDD Cup 2012-Track 2的Tencent CA训练数据集上的实验,测试了方法的有效性.
基金Supported by the National Natural Science Foundation of China (62272214)。
文摘In the era of intelligent economy, the click-through rate(CTR) prediction system can evaluate massive service information based on user historical information, and screen out the products that are most likely to be favored by users, thus realizing customized push of information and achieve the ultimate goal of improving economic benefits. Sequence modeling is one of the main research directions of CTR prediction models based on deep learning. The user's general interest hidden in the entire click history and the short-term interest hidden in the recent click behaviors have different influences on the CTR prediction results, which are highly important. In terms of capturing the user's general interest, existing models paid more attention to the relationships between item embedding vectors(point-level), while ignoring the relationships between elements in item embedding vectors(union-level). The Lambda layer-based Convolutional Sequence Embedding(LCSE) model proposed in this paper uses the Lambda layer to capture features from click history through weight distribution, and uses horizontal and vertical filters on this basis to learn the user's general preferences from union-level and point-level. In addition, we also incorporate the user's short-term preferences captured by the embedding-based convolutional model to further improve the prediction results. The AUC(Area Under Curve) values of the LCSE model on the datasets Electronic, Movie & TV and MovieLens are 0.870 7, 0.903 6 and 0.946 7, improving 0.45%, 0.36% and 0.07% over the Caser model, proving the effectiveness of our proposed model.
基金the National Natural Science Foundation of China(Grants No.91746206 and 72132008)to provide funds for conducting experiments.
文摘In the era of computational advertising,advertising effectiveness can be measured by different metrics at different stages of the sales funnel.In the upper funnel,the click-through rate(CTR,the rate of click per impression)represents the attractiveness of the advertising;the conversion rate(CVR,the rate of conversion per click)in the lower funnel indicates the persuasiveness of the advertising.Achieving higher CTR and CVR may need distinct advertising strategies:improving CTR requires raising more consumers'interest in the ad,which is more beneficial to publishers;boosting CVR needs the ad to inspire more consumers’desire in the product(service),which is more profitable to advertisers.In order to study the performance of advertising texts in terms of the two dimensions and reconcile the two different goals,this paper draws on Speech act theory(SAT)in linguistics to classify advertising texts into three types(i.e.,assertive,expressive,and directive),and analyzes how advertising texts can impact consumer behaviors.We further categorize the above three styles of advertising texts into subjective type(i.e.,expressive and directive)and objective type(i.e.,assertive).Based on a field study,we find that subjective advertising with more personalization leads to a higher CTR,while objective advertising with higher consistency with the brand information results in a higher CVR.The results suggest that firms with different marketing goals should utilize different styles of advertising texts to elicit desirable consumer behaviors during different stages of the sales funnel.
