Tree-Based Real-Time Advertisement Recommendation System in Online Broadcasting

Kang, Seongju; Jeong, Chaeeun; Chung, Kwangsue

doi:10.1109/access.2020.3031925

Cited by 14 publications

(5 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To accurately calculate AC, the EA indicator is introduced, and the calculation is shown in equation 3.8. . rastraction describes the purchasing significance of the artwork, fluctuating between -1 and 1 [17]. When the value is -1, it indicates that the user who purchased the work has extremely low evaluation of it.…”

Section: Building a Recommendation Model Based On User Interest Chara...mentioning

confidence: 99%

Personalized Art Work Recommendation System and Methods Based on User Interest Characteristics and Emotional Preferences

2023

SCPE

View full text Add to dashboard Cite

To familiarize users with their interests and hobbies through online data collection, and improve their experience when browsing art works, research based on K-means algorithm has received widespread attention. However, with the explosive growth of various types of art works, it is difficult to estimate the K value of the K-means algorithm when processing these data. To solve this problem, this research predicts the user behavior of Wink dataset based on K-means algorithm, introduces regularization specified process and emotional precision, and generates fusion algorithm. The study first proposes the concept of similar users and calculates the Pearson correlation coefficient between them to determine their similarity; Then several regularization terms are added to the user group, and the prediction results are obtained by changing the parameters; Further screening of art works clustering categories is to address the issue of slow user startup. Finally, the algorithm studied will be applied to the Wink dataset and the prediction accuracy of the particle swarm optimization algorithm will be tested and compared with the fusion algorithm. A total of 400 experiments are conducted, and the fusion algorithm achieve a prediction accuracy of 392 times, with an accuracy rate of 98.0%; The accuracy of particle swarm optimization algorithm is close to that of fusion algorithm, at 88.2%. The experimental results show that the algorithm model proposed in the study can effectively map the relationship between user interest features, emotional factors, and personalized art recommendation, thereby providing users with a good viewing experience.

show abstract

Section: Building a Recommendation Model Based On User Interest Chara...mentioning

confidence: 99%

Personalized Art Work Recommendation System and Methods Based on User Interest Characteristics and Emotional Preferences

2023

SCPE

View full text Add to dashboard Cite

show abstract

“…For any (s, a) ∈ S × A univ , we have q(s, a) ≤ 1 m . Assumption 1 is mild because in real-world applications, e.g., recommendation systems Fu et al [2021] and online advertising Kang et al [2020], it is often the case that users are only attracted to and click on a few items in a recommended list. Also, in multi-step (e.g., category-based) recommendation, the interests of users usually converge to a single branch in the end (in expectation).…”

Section: Problem Formulationmentioning

confidence: 99%

“…Standard RL Jaksch et al [2010], Agrawal & Jia [2017], Azar et al [2017], Jin et al [2018], Zanette & Brunskill [2019] considers taking only a single action in a state and formulates a single H-step path model. However, in many real-world applications such as recommendation systems Fu et al [2021] and online advertising Kang et al [2020], we often need to select multiple options at a time, and each option can trigger a corresponding successor state. For example, in category-based shopping recommendation Fu et al [2021], the recommendation system often displays a list of main categories at the first step, where each one has a probability to be clicked.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Branching Reinforcement Learning

Du¹,

Chen²

2022

Preprint

View full text Add to dashboard Cite

In this paper, we propose a novel Branching Reinforcement Learning (Branching RL) model, and investigate both Regret Minimization (RM) and Reward-Free Exploration (RFE) metrics for this model. Unlike standard RL where the trajectory of each episode is a single H-step path, branching RL allows an agent to take multiple base actions in a state such that transitions branch out to multiple successor states correspondingly, and thus it generates a tree-structured trajectory. This model finds important applications in hierarchical recommendation systems and online advertising. For branching RL, we establish new Bellman equations and key lemmas, i.e., branching value difference lemma and branching law of total variance, and also bound the total variance by only O(H 2 ) under an exponentially-large trajectory. For RM and RFE metrics, we propose computationally efficient algorithms BranchVI and BranchRFE, respectively, and derive nearly matching upper and lower bounds. Our results are only polynomial in problem parameters despite exponentially-large trajectories.

show abstract

“…In the product recommendation platform, new users register every day, and new products are also on the shelves. e system cannot clarify the interest level of the new products or new users [17,18]. Finally, there is the issue of information expiration.…”

Section: Overall Frameworkmentioning

confidence: 99%

Personalized Product Recommendation Model of Automatic Question Answering Robot Based on Deep Learning

Peng¹,

Xu²

2022

Journal of Robotics

View full text Add to dashboard Cite

The collaborative filtering algorithm widely used in recommendation systems has problems with the sparsity of scoring data and the cold start of new products. A personalized product recommendation model for automated question-answering robots using deep learning is proposed. First, a personalized attention mechanism at the word level and the comment level is proposed, and the comments and users are individually coded. Then, the bidirectional gated recurrent unit (Bi-GRU) is used to construct the score prediction matrix, and through the dynamic collaborative filtering algorithm to integrate the time characteristics of the user’s interest changes. Finally, the feature codes of the users and products are input into the Bi-GRU model for learning, so as to output the recommendation list of personalized products of the automated question answering robot. Experimental results based on the JD and Tianchi datasets show that the training loss of the proposed model is lower than 45 and 23, respectively. And HR@15 and MRR@15 exceed 48 and 15, respectively, which are better than other comparison models. It can better adapt to the actual needs of automatic question-answering robots.

show abstract

Tree-Based Real-Time Advertisement Recommendation System in Online Broadcasting

Cited by 14 publications

References 41 publications

Personalized Art Work Recommendation System and Methods Based on User Interest Characteristics and Emotional Preferences

Personalized Art Work Recommendation System and Methods Based on User Interest Characteristics and Emotional Preferences

Branching Reinforcement Learning

Personalized Product Recommendation Model of Automatic Question Answering Robot Based on Deep Learning

Contact Info

Product

Resources

About