Deep reinforcement learning has shown great potential in improving system performance autonomously, by learning from iterations with the environment. However, traditional reinforcement learning approaches are designed to work in static environments. In many real-world problems, the environments are commonly dynamic, in which the performance of reinforcement learning approaches can degrade drastically. A direct cause of the performance degradation is the high-variance and biased estimation of the reward, due to the distribution shifting in dynamic environments. In this paper, we propose two techniques to alleviate the unstable reward estimation problem in dynamic environments, the stratified sampling replay strategy and the approximate regretted reward, which address the problem from the sample aspect and the reward aspect, respectively. Integrating the two techniques with Double DQN, we propose the Robust DQN method. We apply Robust DQN in the tip recommendation system in Taobao online retail trading platform. We firstly disclose the highly dynamic property of the recommendation application. We then carried out online A/B test to examine Robust DQN. The results show that Robust DQN can effectively stabilize the value estimation and, therefore, improves the performance in this real-world dynamic environment.

Stabilizing Reinforcement Learning in Dynamic Environment with Application to Online Recommendation

In the field of Artificial Intelligence Machine learning provides the automatic systems which learn and improve itself from experience without being explicitly programmed. In this research work a movie recommender system is built using the K-Means Clustering and K-Nearest Neighbor algorithms. The movielens dataset is taken from kaggle. The system is implemented in python programming language. The proposed work deals with the introduction of various concepts related to machine learning and recommendation system. In this work, various tools and techniques have been used to build recommender systems. Various algorithms such as K-Means Clustering, KNN, Collaborative Filtering, Content-Based Filtering have been described in detail. Further, after studying different types of machine learning algorithms, there is a clear picture of where to apply which algorithm in different areas of industries such as recommender systems, e-commerce, etc. Then there is an illustration of how implementations and working of the proposed system are used for the implementation of the movie recommender system. Various building blocks of the proposed system such as Architecture, Process Flow, Pseudo Code, Implementation and Working of the System is described in detail. Finally, in this work for different cluster values, different values of Root Mean Squared Error are obtained. In this proposed work as the no of clusters decreases, the value of RMSE also decreases. The best value of RMSE obtained is 1.081648. The results given by the proposed system are better than the existing technique on the basis of RMSE value.

Movie Recommender System Using K-Means Clustering AND K-Nearest Neighbor

Today’s Recommender system is a relatively new area of research in machine learning. The recommender system’s main idea is to build relationship between the products, users and make the decision to select the most appropriate product to a specific user. There are four main ways that recommender systems produce a list of recommendations for a user – content-based, Collaborative, Demographic and hybrid filtering. In content-based filtering the model uses specifications of an item in order to recommend additional items with similar properties. Collaborative filtering uses past behavior of the user like items that a user previously viewed or purchased, In summation to any ratings the user gave those items rate and similar conclusions made by other user’s items list. To predicts items that the user may find interesting. Demographic filtering is view user profile data like age category, gender, education and living area to find similarities with other profiles to get a new recommender list. Hybrid filtering combines all three filtering techniques. This paper introduces survey about recommendation systems, techniques, challenges the face recommender systems and list some research papers solve these challenges.

Recommender Systems Challenges and Solutions Survey

Internet of Things (IoT) has emerged in many industries, such as health care, transportation, agriculture, manufacturing, smart homes, to name a few. It paves the path for massive applications on the user level to enhance the quality of life or service, and on the decision-makers’ level to provide a sustainable increase in revenue. IoT principally connects different physical objects (e.g., sensors) and enables them to communicate, collect, and share data. In the Era of IoT, Recommendation systems provide personalized recommendations based on the user's historical datasets collected from the IoT devices. These recommendations enable an efficient decision-making process by suggesting relevant products, resources, and information. This paper provides an overview of various multi-layers IoT architectures, and IoT-based recommendation systems with an emphasis on their advantages, disadvantages, application domains, and validation metrics for quality assessment.

IoT-based Recommendation Systems – An Overview

This study presents a configurable approach for recommendations which determines the suitable recommendation method for each field based on the characteristics of its data, the method includes determining the suitable technique for selecting a representative sample of the provided data. Then selecting the suitable feature weighting measure to provide a correct weight for each feature based on its effect on the recommendations. Finally, selecting the suitable algorithm to provide the required recommendations. The proposed configurable approach could be applied on different domains. The experiments have revealed that the approach is able to provide recommendations with only 0.89 error rate percentage.

A proposed configurable approach for recommendation systems via data mining techniques

Recommendation systems have become extremely common in recent years. It helps the customer to discover information and settle on choices where they do not have the required learning to judge a specific item. It can be utilized as a part of different diverse approaches to encourage its customer with effective information sorting. It is a software tool and techniques that provide suggestion based on the customer's taste to discover new appropriate thing for them by filtering personalized information based on the user's preferences from a large volume of information. Users taste and preferences should be constructed accurately in order to provide most relevant suggestions. This survey paper compare's and details the various type of recommender system and popular recommendation algorithms and its uses.

/pdf/a-survey-on-recommendation-system-3mmtih9ula.pdf

A Survey on Recommendation System

Nowadays, we are living in an age recommendation, but the proper recommendation needs more accurate and relevant datas as their inputs. Rating databases like MovieLence or Netflix have long been popular and being widely used in recommendation system areas for research in past decades. But nowadays, they become irrelevant due to lack of new and relevant datas. Nowadays, social media like Facebook and Twitter become the most popular for researchers due to availability of large amount of new and relevant datas. In this paper, we have built a recommendation engine by analyzing rating datasets collected from Twitter to recommend movies to specific user using R.

Personalized Movie Recommendation System Using Twitter Data

Dramatic growth of social media has created remarkable interest among Internet users nowadays. Information from these Web sites in the form of reviews, feedbacks, ratings, etc., can be utilized for various purposes like to find out users’ taste or interest to develop a proper marketing strategy, maybe for a survey about the product by using sentiment analysis. Twitter is generally used for posting long comments in short status. Twitter offers organizations a fast and powerful approach to investigate customers’ viewpoints toward the critical to success in the open market. Previously we calculate sentiment of each word for the sentiment, which may or may not be accurate because may be the same word used in past for negative review, but presently it is used for positive sense. We propose a method by applying both log function and N-gram techniques to find out the sentiment of the Twitter data in R to build a robust engine to achieve more accuracy.

Debashis Das

Papers

A Survey on Recommendation System

Personalized Movie Recommendation System Using Twitter Data

Sentiment Analysis Using N-gram Technique