A New Contextual Influencer User Measure to Improve the Accuracy of Recommender System

2018 ◽  
Vol 9 (4) ◽  
pp. 38-51
Author(s):  
Maryam Jallouli ◽  
Sonia Lajmi ◽  
Ikram Amous
Keyword(s):  
Social Network ◽  
Recommender System ◽  
Recommendation System ◽  
Mean Absolute Error ◽  
Absolute Error ◽  
User Model ◽  
Context Aware ◽  
Additional Information ◽  
Context Element ◽  
Cold Start Problem

In the last decade, social-based recommender systems have become the best way to resolve a user's cold start problem. In fact, it enriches the user's model by adding additional information provided from his social network. Most of those approaches are based on a collaborative filtering and compute similarities between the users. The authors' preliminary objective in this work is to propose an innovative context aware metric between users (called contextual influencer user). These new similarities are called C-COS, C-PCC and C-MSD, where C refers to the category. The contextual influencer user model is integrated into a social based recommendation system. The category of the items is considered as the most pertinent context element. The authors' proposal is implemented and tested within the food dataset. The experimentation proved that the contextual influencer user measure achieves 0.873, 0.874, and 0.882 in terms of Mean Absolute Error (MAE) corresponding to C-cos, C-pcc and C-msd, respectively. The experimental results showed that their model outperforms several existing methods.

Hybrid Model for Movie Recommendation System Using Fireflies and Fuzzy C-Means

2019 ◽  
Vol 11 (2) ◽  
pp. 1-13
Author(s):  
M. Sandeep Kumar ◽  
Prabhu J.
Keyword(s):  
Hybrid Model ◽  
Recommender System ◽  
Firefly Algorithm ◽  
Recommendation System ◽  
Mean Absolute Error ◽  
Research Work ◽  
Absolute Error ◽  
Experimental Result ◽  
Fuzzy C Means ◽  
Fcm Clustering

In the era of Big Data, extremely complicated data is delivered from the system, of which it is impossible to collect the correct information with an online platform. In this research work, it provides a hybrid model for a movie-based recommender system; based on meta-heuristic firefly algorithm and fuzzy c-means (FCM) clustering technique to evaluate rating of a movie for a specific user based on the similarity of users and historical data. The firefly algorithm was employed in the movie lens dataset to get the initial cluster and also to initialize the position of clusters. FCM is used to classify the similarity of the user ratings. The proposed collaborative recommender system performed well regarding accuracy and precision. Various metrics are used in a movie lens dataset like mean absolute error (MAE), precision, and recall. The experimental result delivered by the system provides more efficient performance compared to the existing system in term of mean absolute error (MAE).

scholarly journals Pembuatan Modul Rekomendasi pada OpenCart Menggunakan Metode Item-Based Collaborative Filtering

2013 ◽  
Vol 2 (1) ◽  
pp. 9
Author(s):  
Kirana Nuryunita ◽  
Yani Nurhadryani
Keyword(s):  
Collaborative Filtering ◽  
Recommender System ◽  
Management System ◽  
Mean Absolute Error ◽  
Absolute Error ◽  
Content Management ◽  
Cosine Similarity ◽  
Weighted Sum ◽  
Content Management System

<p>Penelitian ini bertujuan menambahkan modul rekomendasi pada content management system Opencart. Salah satu pendekatan dalam rekomendasi adalah item-based collaborative filtering. Metode item-based collaborative filtering dapat mengurangi waktu eksekusi perhitungan. Metode item-based collaborative filtering pada penelitian ini menggunakan perhitungan adjusted cosine similarity untuk menghitung nilai kemiripan antarbuku dan weighted sum untuk menghitung nilai prediksi rate buku. Untuk mendapatkan rekomendasi, pengguna harus melakukan login dan memberikan rate pada buku. Berdasarkan rate pengguna, nilai kemiripan dihitung menggunakan adjusted cosine similarity. Berdasarkan kemiripan antarbuku, nilai prediksi rate buku dicari menggunakan weighted sum. Sebelum buku direkomendasikan kepada pengguna, kategori prediksi buku dicocokkan dengan kategori buku yang telah diberi rate oleh pengguna. Penelitian ini menggunakan 300 buku dan 30 pengguna sebagai data. Dari hasil penelitian, hanya 17 pengguna yang mendapatkan rekomendasi. Pengujian dilakukan dengan menganalisis waktu eksekusi dan keakuratan rekomendasi. Waktu eksekusi dalam pengujian ini adalah 1.60 detik. Untuk menghitung keakuratan rekomendasi, penelitian ini menggunakan mean absolute error dengan hasil perhitungan 0.15.</p><p>Kata kunci: e-commerce, item-based collaborative filtering, recommender system.</p>

scholarly journals Graph Neural Network and Context-Aware Based User Behavior Prediction and Recommendation System Research

2020 ◽  
Vol 2020 ◽  
pp. 1-14
Author(s):  
Qian Gao ◽  
Pengcheng Ma
Keyword(s):  
Neural Network ◽  
Recommendation System ◽  
User Behavior ◽  
Absolute Error ◽  
User Preferences ◽  
Feature Interaction ◽  
Context Aware ◽  
Behavior Prediction ◽  
Complex Interactions ◽  
System Maps

Due to the influence of context information on user behavior, context-aware recommendation system (CARS) has attracted extensive attention in recent years. The most advanced context-aware recommendation system maps the original multi-field features into a shared hidden space and then simply connects it to a deep neural network (DNN) or other specially designed networks. However, for different areas, the ability of modeling complex interactions in a sufficiently flexible and explicit way is limited by the simple unstructured combination of feature fields. Therefore, it is hard to get the accurate results of the user behavior prediction. In this paper, a graph structure is used to establish the interaction between context and users/items. Through modeling user behavior, we can explore user preferences in different context environments, so as to make personalized recommendations for users. In particular, we construct a context-user and context-item interactions graph separately. In the interactions graph, each node is composed of a user feature field, an item feature field, and a feature field of different contexts. Different feature fields can interact through edges. Therefore, the task of modeling feature interaction can be transformed into modeling the node interaction on the corresponding graph. To this end, an innovative model called context-aware graph neural network (CA-GNN) model is designed. Furthermore, in order to obtain more accurate and efficient recommendation results, first, we innovatively use the attention mechanism to improve the interpretability of CA-GNN; second, we innovatively use the degree of physical fatigue features which has never been used in traditional CARS as critical contextual feature information into our CA-GNN. We simulated the Food and Yelp datasets. The experimental results show that CA-GNN is better than other methods in terms of root mean square error (RMSE) and mean absolute error (MAE).

Research on Recommendation Algorithm in Social Networks

2014 ◽  
Vol 496-500 ◽  
pp. 1865-1868
Author(s):  
Hu Xin Tang ◽  
Xu Qian
Keyword(s):  
Social Networks ◽  
Social Network ◽  
Recommender System ◽  
Recommendation System ◽  
Simulation Experiment ◽  
Recommendation Algorithm ◽  
Simulation Experiments ◽  
Correlation Algorithm ◽  
Rapid Calculation ◽  
Relative Attractiveness

Research status and development of the recommendation system are studied, the focus of evaluation of recommender system and recommender system based on social network in two aspects, and puts forward some improved algorithm, and achieved certain results. KDD Cup 2012 Track data for the simulation experiments on the correlation algorithm based on search engine, has been shown in different positions on the relative attractiveness of advertising, numerical user. At the same time, rapid calculation of the degree of correlation between a user and other users of an algorithm is given, and then quickly given the recommendation results. KDD Cup 2012 Track data for the simulation experiment of the algorithm, and the analysis result is given.

scholarly journals An AI-Based Exercise Prescription Recommendation System

2021 ◽  
Vol 11 (6) ◽  
pp. 2661
Author(s):  
Hung-Kai Chen ◽  
Fueng-Ho Chen ◽  
Shien-Fong Lin
Keyword(s):  
Recommendation System ◽  
Exercise Prescription ◽  
Rest Heart Rate ◽  
Mean Absolute Error ◽  
Absolute Error ◽  
Exercise Mode ◽  
Preventive Cardiology ◽  
Community Data ◽  
Testing Accuracy ◽  
Digital Training

The European Association of Preventive Cardiology Exercise Prescription in Everyday Practice and Rehabilitative Training (EXPERT) tool has been developed for digital training and decision support in cardiovascular disease patients in clinical practice. Exercise prescription recommendation systems for sub-healthy people are essential to enhance this dominant group’s physical ability as well. This study aims to construct a guided exercise prescription system for sub-healthy groups using exercise community data to train an AI model. The system consists of six modules, including three-month suggested exercise mode (3m-SEM), predicted value of rest heart rate (rest HR) difference after following three-month suggested exercise mode (3m-PV), two-month suggested exercise mode (2m-SEM), predicted value of rest HR difference after following two-month suggested exercise mode (2m-PV), one-month suggested exercise mode (1m-SEM) and predicted value of rest HR difference after following one-month suggested exercise mode (1m-PV). A new user inputs gender, height, weight, age, and current rest HR value, and the above six modules will provide the user with a prescription. A four-layer neural network model is applied to construct the above six modules. The AI-enabled model produced 95.80%, 100.00%, and 95.00% testing accuracy in 1m-SEM, 2m-SEM, and 3m-SEM, respectively. It reached 3.15, 2.89, and 2.75 BPM testing mean absolute error in 1m-PV, 2m-PV, and 3m-PV. The developed system provides quantitative exercise prescriptions to guide the sub-healthy group to engage in effective exercise programs.

scholarly journals Location Based Agricultural Product Recommendation System Using Novel KNN Algorithm

2019 ◽  
pp. 945-950
Author(s):  
Sachin J ◽  
Geethatharani P ◽  
Surya M K ◽  
Kavin K V
Keyword(s):  
Euclidean Distance ◽  
Recommendation System ◽  
Mean Absolute Error ◽  
Market Price ◽  
Absolute Error ◽  
Agricultural Products ◽  
Agricultural Product ◽  
Product Recommendation ◽  
Use Of Data ◽  
Great Circle Distance

It is evident that the need for personalized product recommendation is much needed these days. Generally, product recommender systems are implemented in web servers that make use of data, implicitly obtained as results of the collection of Web browsing patterns of the users. Here, the project's motive is to provide location-based agricultural product recommendation system using a novel KNN algorithm by ensuring effective communication and transparency in agriculture trade marketing among buyers and sellers (farmers). It helps the farmer to fix up the market price by preventing the rue pricing of their products. The farmer can post their products into the application with price and other details like a timestamp of harvesting, color, size, the absence of pest, freshness, ripeness etc. Based on the location, the distance between the seller and buyer is calculated using great circle distance. An improved Novel KNN algorithm is used to find the K Nearest Seller by calculating the distance between the sellers and buyer using a Euclidean distance metric. The details posted by the farmers and buyers are stored and updated in a database dynamically. The recommender system recommends nearest sellers and their agricultural products based on buyer interest. The performance of the system is analyzed in terms of accuracy and mean absolute error.

scholarly journals Conversational Recommender System with Explanation Facility Using Semantic Reasoning

2016 ◽  
Vol 2 (1) ◽  
pp. 1
Author(s):  
Zk Abdurahman Baizal ◽  
Nur Rahmawati
Keyword(s):  
Recommender System ◽  
Recommendation System ◽  
User Model ◽  
Functional Requirements ◽  
Inference Method ◽  
Semantic Reasoning ◽  
Filtering Method ◽  
Knowledge Based ◽  
The Relationship ◽  
Number Of Iterations

<p>Conversational recommender system is system that provides dialogue as user guide to obtain information from the user, in order to obtain preference for products needed. This research implements conversational recommender system with knowledge-based in the smartphone domain with an explanation facility. The recommended products are obtained based on the functional requirements of the user. Therefore, this study use ontology model as a knowledge to be more flexible in finding products that is suitable with the functional requirements of the user that is by tracing a series of semantic based on relationships in order to obtain the user model. By exploiting the relationship between instances of user models, the explanation facility generated can be more natural. Our filtering method uses semantic reasoning with inference method to avoid overspecialization. The evaluation show that the performance of our recommender system with explanation facilities is more efficient than the recommendation system without explanation facility, that can be seen from the number of iterations. We also notice that our system has accuracy of 84%.</p>

scholarly journals Building A Movie Recommendation System Using Collaborative Filtering With TF-IDF

2021 ◽  
Vol 9 (9) ◽  
pp. 181-191
Author(s):  
S. A. Azeem Farhan
Keyword(s):  
Collaborative Filtering ◽  
Language Processing ◽  
Recommender System ◽  
Recommendation System ◽  
Mean Squared Error ◽  
Information Filtering ◽  
Recommendation Systems ◽  
Absolute Error ◽  
Machine Learning Algorithms ◽  
Value Decomposition

Abstract: The recommendation problem involves the prediction of a set of items that maximize the utility for users. As a solution to this problem, a recommender system is an information filtering system that seeks to predict the rating given by a user to an item. There are theree types of recommendation systesms namely Content based, Collaborative based and the Hybrid based Recommendation systems. The collaborative filtering is further classified into the user based collaborative filtering and item based collaborative filtering. The collaborative filtering (CF) based recommendation systems are capable of grasping the interaction or correlation of users and items under consideration. We have explored most of the existing collaborative filteringbased research on a popular TMDB movie dataset. We found out that some key features were being ignored by most of the previous researches. Our work has given significant importance to 'movie overviews' available in the dataset. We experimented with typical statistical methods like TF-IDF , By using tf-idf the dimensions of our courps(overview and other text features) explodes, which creates problems ,we have tackled those problems using a dimensionality reduction technique named Singular Value Decomposition(SVD). After this preprocessing the Preprocessed data is being used in building the models. We have evaluated the performance of different machine learning algorithms like Random Forest and deep neural networks based BiLSTM. The experiment results provide a reliable model in terms of MAE(mean absolute error) ,RMSE(Root mean squared error) and the Bi-LSTM turns out to be a better model with an MAE of 0.65 and RMSE of 1.04 ,it generates more personalized movie recommendations compared to other models. Keywords: Recommender system, item-based collaborative filtering, Natural Language Processing, Deep learning.

Sign in / Sign up

Export Citation Format

Share Document