scholarly journals Hybrid model for movie recommendation system using content K-nearest neighbors and restricted Boltzmann machine

2021 ◽  
Vol 23 (1) ◽  
pp. 445
Author(s):  
Dayal Kumar Behera ◽  
Madhabananda Das ◽  
Subhra Swetanisha ◽  
Prabira Kumar Sethy
Keyword(s):  
Collaborative Filtering ◽  
Hybrid Model ◽  
Recommendation System ◽  
Nearest Neighbors ◽  
Restricted Boltzmann Machine ◽  
Boltzmann Machine ◽  
K Nearest Neighbors ◽  
Benchmark Datasets ◽  
Data Content ◽  
Rating Prediction

<span>One of the most commonly used techniques in the recommendation framework is collaborative filtering (CF). It performs better with sufficient records of user rating but is not good in sparse data. Content-based filtering works well in the sparse dataset as it finds the similarity between movies by using attributes of the movies. RBM is an energy-based model serving as a backbone of deep learning and performs well in rating prediction. However, the rating prediction is not preferable by a single model. The hybrid model achieves better results by integrating the results of more than one model. This paper analyses the weighted hybrid CF system by integrating content K-nearest neighbors (KNN) with restricted Boltzmann machine (RBM). Movies are recommended to the active user in the proposed system by integrating the effects of both content-based and collaborative filtering. Model efficacy was tested with MovieLens benchmark datasets.</span>

scholarly journals Discretization Based Framework to Improve the Recommendation Quality

2021 ◽  
Vol 18 (3) ◽  
Author(s):  
Bilal Ahmed ◽  
Wang Li
Keyword(s):  
Collaborative Filtering ◽  
Recommendation System ◽  
Information Filtering ◽  
Nearest Neighbors ◽  
User Preferences ◽  
Primary Concern ◽  
K Nearest Neighbors ◽  
Chi Square ◽  
Similarity Computation ◽  
Rating Prediction

Recommendation systems are information filtering software that delivers suggestions about relevant stuff from a massive collection of data. Collaborative filtering approaches are the most popular in recommendations. The primary concern of any recommender system is to provide favorable recommendations based on the rating prediction of user preferences. In this article, we propose a novel discretization based framework for collaborative filtering to improve rating prediction. Our framework includes discretization-based preprocessing, chi-square based attribution selection, and K-Nearest Neighbors (KNN) based similarity computation. Rating prediction affords some basis for the judgment to decide whether recommendations are generated or not, subject to the ratio of performance of any recommendation system. Experiments on two datasets MovieLens and BookCrossing, demonstrate the effectiveness of our method.

scholarly journals Performa Algoritma User K-Nearest Neighbors pada Sistem Rekomendasi di Tokopedia

2020 ◽  
Vol 5 (3) ◽  
pp. 302
Author(s):  
Rama Dian Syah
Keyword(s):  
Recommendation System ◽  
Mean Absolute Error ◽  
Prediction Method ◽  
Large Data ◽  
Nearest Neighbors ◽  
Absolute Error ◽  
K Nearest Neighbors ◽  
Algorithm Performance ◽  
Specific Product ◽  
Rating Prediction

The biggest marketplace in Indonesia such as Tokopedia has data on e-commerce activities that always increase with time. Large data growth in Marketplace can cause problems for users. Buyers who have difficulty in finding the best product that suits their needs and sellers who have difficulty in promoting products that are often visited by buyers can be overcome. The recommendation system can overcome these problems by providing specific product recommendations to be promoted and offered to buyers. This research implements the Recommendation System using the Item Rating Prediction Method by applying the User K-Nearest Neighbors Algorithm. The Recommendation System provides recommendations based on ratings on products given by the buyer. Algorithm performance in Recommendation System is measured by the parameters of Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Normalized Mean Absolute Error (NMAE). The performance values obtained are RMSE = 0.713, MAE = 0.488 and NMAE = 0.122. Perfomance values below 1 proves that the User K-Nearest Neighbors Algorithm is suitable as a rating prediction model on recommendation system.

On identifying k -nearest neighbors in neighborhood models for efficient and effective collaborative filtering

2018 ◽  
Vol 278 ◽  
pp. 134-143 ◽  
Author(s):  
Dong-Kyu Chae ◽  
Sang-Chul Lee ◽  
Si-Yong Lee ◽  
Sang-Wook Kim
Keyword(s):  
Collaborative Filtering ◽  
Nearest Neighbors ◽  
K Nearest Neighbors ◽  
Neighborhood Models

scholarly journals Voiceprint Identification for Limited Dataset Using the Deep Migration Hybrid Model Based on Transfer Learning

Sensors ◽  
2018 ◽  
Vol 18 (7) ◽  
pp. 2399 ◽  
Author(s):  
Cunwei Sun ◽  
Yuxin Yang ◽  
Chang Wen ◽  
Kai Xie ◽  
Fangqing Wen
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Transfer Learning ◽  
Hybrid Model ◽  
Data Augmentation ◽  
Small Sample ◽  
Restricted Boltzmann Machine ◽  
Small Samples ◽  
Boltzmann Machine ◽  
Training Time

The convolutional neural network (CNN) has made great strides in the area of voiceprint recognition; but it needs a huge number of data samples to train a deep neural network. In practice, it is too difficult to get a large number of training samples, and it cannot achieve a better convergence state due to the limited dataset. In order to solve this question, a new method using a deep migration hybrid model is put forward, which makes it easier to realize voiceprint recognition for small samples. Firstly, it uses Transfer Learning to transfer the trained network from the big sample voiceprint dataset to our limited voiceprint dataset for the further training. Fully-connected layers of a pre-training model are replaced by restricted Boltzmann machine layers. Secondly, the approach of Data Augmentation is adopted to increase the number of voiceprint datasets. Finally, we introduce fast batch normalization algorithms to improve the speed of the network convergence and shorten the training time. Our new voiceprint recognition approach uses the TLCNN-RBM (convolutional neural network mixed restricted Boltzmann machine based on transfer learning) model, which is the deep migration hybrid model that is used to achieve an average accuracy of over 97%, which is higher than that when using either CNN or the TL-CNN network (convolutional neural network based on transfer learning). Thus, an effective method for a small sample of voiceprint recognition has been provided.

Sign in / Sign up

Export Citation Format

Share Document