Abstract:Combining a deep neural network with fuzzy theory, this paper proposes an advertising click-through rate (CTR) prediction approach based on a fuzzy deep neural network (FDNN). In this approach, fuzzy Gaussian-Bernoulli restricted Boltzmann machine (FGBRBM) is first applied to input raw data from advertising datasets. Next, fuzzy restricted Boltzmann machine (FRBM) is used to construct the fuzzy deep belief network (FDBN) with the unsupervised method layer by layer. Finally, fuzzy logistic regression (FLR) is u… Show more
“…Fuzzy system has different characteristics. Fuzzy system can summarize and express corresponding knowledge based on human experience, which is more convenient to understand [ 12 – 14 ]. In fuzzy system, knowledge is stored in the rule set, and the number of rules can be controlled and adjusted, which is less than the calculation of neural network [ 15 ].…”
In recent years, with the gradual development of sports, the competition between athletes is becoming more and more fierce. The long training time and heavy body load of athletes lead to the increase of the incidence of sports injury, and the evaluation and analysis of athletes’ sports injury need a lot of manpower and material resources. In order to improve the calculation efficiency of sports injury estimation results and save the cost of estimation and analysis, we propose a sports injury estimation model based on the algorithm model of mutation fuzzy neural network. The sports injury model constructed in this paper can not only systematically evaluate and analyze the degree of sports injury of athletes, but also improve the accuracy and efficiency; at the same time, it has universality for the evaluation and analysis of the degree of sports injury. The construction of this model provides the theoretical basis of big data algorithm for the prevention of sports injury and the application of mutation fuzzy neural network in the field of sports.
“…Fuzzy system has different characteristics. Fuzzy system can summarize and express corresponding knowledge based on human experience, which is more convenient to understand [ 12 – 14 ]. In fuzzy system, knowledge is stored in the rule set, and the number of rules can be controlled and adjusted, which is less than the calculation of neural network [ 15 ].…”
In recent years, with the gradual development of sports, the competition between athletes is becoming more and more fierce. The long training time and heavy body load of athletes lead to the increase of the incidence of sports injury, and the evaluation and analysis of athletes’ sports injury need a lot of manpower and material resources. In order to improve the calculation efficiency of sports injury estimation results and save the cost of estimation and analysis, we propose a sports injury estimation model based on the algorithm model of mutation fuzzy neural network. The sports injury model constructed in this paper can not only systematically evaluate and analyze the degree of sports injury of athletes, but also improve the accuracy and efficiency; at the same time, it has universality for the evaluation and analysis of the degree of sports injury. The construction of this model provides the theoretical basis of big data algorithm for the prevention of sports injury and the application of mutation fuzzy neural network in the field of sports.
“…In the real production environment, an advertising network forecasting application does require the prediction results within milliseconds or even microseconds [37]. Therefore, the approach to the short interval prediction using Neural Network is not just scientifically important but also practically required since it does take great advantages in fast forecasting.…”
In the field of advertising technology, it is a key task to forecast posterior click distribution since 66% of advertising transactions depend on cost per click model. However, due to the General Data Protection Regulation, machine learning techniques to forecast posterior click distribution based on the sequences of an identified user's actions are restricted in European countries. To overcome this barrier, we introduce a contextual behavior concept for the advertising network environment and propose a new hybrid model, which we call the Long Short Term Memory-Hawkes model by combining a stochastic-based generative model and a machine learning-based predictive model. Also, to meet the computational efficiency for the heavy demand in mobile advertisement market, we define gradient exponential kernel with just three hyper parameters to minimize residuals. We have carefully tested our proposed model with production data and found that the LSTM-Hawkes model reduces the Mean Squared Error by at least 27.1% and up to 83.8% on average in comparison to the existing Hawkes Process based algorithm, Hawkes Intensity Process, as well as 39.77% on average in comparison to Multivariate Linear Regression. We have also found that our proposed model improves the forecast accuracy by about 21.2% on average.
“…A high CTR means the targeting is effective. Jiang et al (2018) explain a high CTR means that the marketing campaign is contributing to more leads and sales. On the other hand, a low CTR means the ads are not appropriate for the audience.…”
Ad effectiveness on online platforms including Facebook and Google is a challenge for businesses. Since, ad networks use algorithms, ad effectiveness as measured by CTRs is not well understood by marketing and sales executives. CTR prediction with deep neural networks can improve ad CTRs. The AI solution for ad CTRs is useful across industry sectors. In the solution, RNN Models learn representations of sequences by maintaining internal states which are updated sequentially and used as proxy for target prediction. Evidence from research shows that deep neural networks could help businesses enhance CTRs.
Problem description:Ads have become cost-intensive for businesses across industry sectors. Since, businesses operate in competitive environments, it is critical for the executive management to focus on good decision-making. Regardless of the business sector, it is critical to enhance the decision-making process as this could mean a significant reduction in ad spending. Machine learning and AI could enhance the ad effectiveness decision-making. Smith (2018) argues that several factors should be considered while analysing whether Click-Through Rate (CTR) is good. Since, achieving higher CTRs is challenging to accomplish (Cheng et al., 2012), machine learning and AI could be deployed for improving CTR. Markus (2017) explains how Google and Facebook use CTRs to determine the quality of ads. Each ad is given a quality score which is based on the CTR. Ads use bidding systems which consider the quality score for ad placement. A high-quality score could outrank the competition and do away the need for outbidding the competition. This means the ad shows higher up in the user's feed at a lower cost-per-click. Thus, the challenge for advertising is to increase the CTRs. Strong CTRs could be achieved by good targeting and creativity. Several factors come into play to make a good CTR. Ad positioning, imagery quality and keywords are the main factors that impact CTR. Though these guidelines are useful there is no clear solution for determining what is a good CTR.Literature review: This section has a review of relevant literature on ad effectiveness, the most significant problem for online ad spending. The discussion shows the relevance of the metric CTR and how AI could solve this problem for businesses based on evidence from academia and industry experts.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.