A deep learning based algorithm for multi-criteria recommender systems

Shambour, Qusai

doi:10.1016/j.knosys.2020.106545

Cited by 86 publications

(26 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, deep learning-based recommender systems have also obtained very promising results [12,13]. Furthermore, although IoT is considered a key concept in smart tourism, it is very rarely applied in smart tourism recommendation systems [14].…”

Section: Introductionmentioning

confidence: 99%

Deep learning and Internet of Things for tourist attraction recommendations in smart cities

Cepeda-Pacheco

Domingo

2022

Neural Comput & Applic

View full text Add to dashboard Cite

We propose a tourist attraction IoT-enabled deep learning-based recommendation system to enhance tourist experience in a smart city. Travelers will enter details about their travels (traveling alone or with a companion, type of companion such as partner or family with kids, traveling for business or leisure, etc.) as well as user side information (age of the traveler/s, hobbies, etc.) into the smart city app/website. Our proposed deep learning-based recommendation system will process this personal set of input features to recommend the tourist activities/attractions that best fit his/her profile. Furthermore, when the tourists are in the smart city, content-based information (already visited attractions) and context-related information (location, weather, time of day, etc.) are obtained in real time using IoT devices; this information will allow our proposed deep learning-based tourist attraction recommendation system to suggest additional activities and/or attractions in real time. Our proposed multi-label deep learning classifier outperforms other models (decision tree, extra tree, k-nearest neighbor and random forest) and can successfully recommend tourist attractions for the first case [(a) searching for and planning activities before traveling] with the loss, accuracy, precision, recall and F1-score of 0.5%, 99.7%, 99.9%, 99.9% and 99.8%, respectively. It can also successfully recommend tourist attractions for the second case [(b) looking for activities within the smart city] with the loss, accuracy, precision, recall and F1-score of 3.7%, 99.5%, 99.8%, 99.7% and 99.8%, respectively.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep learning and Internet of Things for tourist attraction recommendations in smart cities

Cepeda-Pacheco

Domingo

2022

Neural Comput & Applic

View full text Add to dashboard Cite

show abstract

“…ECoRec [21], ISVDR [22], DGR-EGB [23], SVM [24], CF-K-means [25], DDL-PMF [27], DPM [15], DPMI [15], DPMF [15] and DIPMI frameworks. Here, the items from the Trip Advisor and Amazon datasets are considered to reorganize and suggest the items to the users according to their rating behavior.…”

Section: Resultsmentioning

confidence: 99%

“…An enhanced ride-sharing framework [24] using Support Vector Machine (SVM) was presented to predict the behavior of new riders by learning the client's opinion after completing their trip. A novel intelligent recommender system was designed [25] International Journal of Intelligent Engineering and Systems, Vol. 15…”

Section: Literature Surveymentioning

confidence: 99%

Recommender System using Distributed Improved Predictive Framework with Matrix Factorization and Random Forest

2022

IJIES

View full text Add to dashboard Cite

Online digital marketing achieves their revenue according to their advertisements or sales assignment when companies have the profitable attention for recommending their products to customers via ranking them. Online customers are not able to guarantee that the items delivered through the recommendation by big data are either comprehensive or applicable to their essentials. In the past few years, recommendation frameworks were broadly applied to analyze the massive amount of data. Among those, a Distributed Predictive model with Matrix factorization and random Forest (DPMF) has achieved high efficiency to predict the item ratings. However, it accounts only for user preferences and opinions whereas the other contextual data are necessary for enhancing the efficiency of rating prediction. In this article, a Distributed Improved Predictive model with a Matrix factorization and random Forest (DIPMF) framework is proposed that considers the elements of social context and the dynamic characteristic of every user for every item to enhance the quality of prediction. The primary aim is to combine the information from the preferences, opinions and social context of each user. The social context of users is multiple features of the context such as differences in the current opinion with earlier opinion, behavior, relationship and interaction. Since each user is connected through relations and interactions. At first, the training dataset is split into an optimal amount of splits for accelerating the parallel and distributed training process. Then, the training process is carried out by the DPMF with the Distributed Improved Predictive model with Matrix factorization-Improved variant (DIPMI) to create the representation of every user's preferences, opinions and social contexts in the training set. Further, the prediction of rating is formulated as a regression challenge and solved via the Random Forest (RF) algorithm that predicts the customer's rating behavior with their opinions and social context for every product. Finally, the experiments are conducted on trip advisor and Amazon datasets to evaluate the efficiency of DIPMI and DIPMF compared to the state-of-the-art recommendation frameworks. The findings exhibit that the DIPMF on the trip advisor dataset achieves an average of 0.6826 Root Mean Square Error (RMSE), 0.5925 Mean Absolute Error (MAE), 0.8369 recommendation quality and 0.0023 Confidence Range (CR) 95% compared to the other frameworks. Similarly, the DIPMF on Amazon dataset achieves an average of 0.7591 RMSE, 0.5704 MAE, 0.8298 recommendation quality and 0.0032 CR 95% compared to the other frameworks.

show abstract

“…After the momentums are computed, these initial biases are corrected by the bias-correction method. Finally, the current weight w t is updated by the update rule of the GD shown in Equation (6). From Algorithm 1, we can find that the first momentum represents the most promising search direction determined by the past and current gradients.…”

Section: Optimization Methods To Train Cnnsmentioning

confidence: 99%

“…In particular, modern neural network models are consisted of deeper layers and more weights than traditional ones to maximize their performance. Accordingly, the latest deep learning models have shown notable abilities in many real-world applications, for example, computer visions (CV) [1,2], data analysis [3,4], personalized services [5,6], internet of things (IoT) [7,8], and natural language processing (NLP) [9,10], et al Among them, particularly, the CV task involving image classification and image semantic segmentation is one of the applications in which the deep learning models have been most actively used. Accordingly, many studies to improve the image processing ability of CNNs are being actively conducted.…”

Section: Introductionmentioning

confidence: 99%

HyAdamC: A New Adam-Based Hybrid Optimization Algorithm for Convolution Neural Networks

Kim

Choi

2021

Sensors

View full text Add to dashboard Cite

As the performance of devices that conduct large-scale computations has been rapidly improved, various deep learning models have been successfully utilized in various applications. Particularly, convolution neural networks (CNN) have shown remarkable performance in image processing tasks such as image classification and segmentation. Accordingly, more stable and robust optimization methods are required to effectively train them. However, the traditional optimizers used in deep learning still have unsatisfactory training performance for the models with many layers and weights. Accordingly, in this paper, we propose a new Adam-based hybrid optimization method called HyAdamC for training CNNs effectively. HyAdamC uses three new velocity control functions to adjust its search strength carefully in term of initial, short, and long-term velocities. Moreover, HyAdamC utilizes an adaptive coefficient computation method to prevent that a search direction determined by the first momentum is distorted by any outlier gradients. Then, these are combined into one hybrid method. In our experiments, HyAdamC showed not only notable test accuracies but also significantly stable and robust optimization abilities when training various CNN models. Furthermore, we also found that HyAdamC could be applied into not only image classification and image segmentation tasks.

show abstract

A deep learning based algorithm for multi-criteria recommender systems

Cited by 86 publications

References 21 publications

Deep learning and Internet of Things for tourist attraction recommendations in smart cities

Deep learning and Internet of Things for tourist attraction recommendations in smart cities

Recommender System using Distributed Improved Predictive Framework with Matrix Factorization and Random Forest

HyAdamC: A New Adam-Based Hybrid Optimization Algorithm for Convolution Neural Networks

Contact Info

Product

Resources

About