RMDL: Random Multimodel Deep Learning for Classification

3 May 2018Kamran KowsariMojtaba HeidarysafaDonald E. BrownKiana Jafari MeimandiLaura E. Barnes

The continually increasing number of complex datasets each year necessitates ever improving machine learning methods for robust and accurate categorization of these data. This paper introduces Random Multimodel Deep Learning (RMDL): a new ensemble, deep learning approach for classification... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Document Classification 20NEWS RMDL Accuracy 87.91 # 1
Multi-Label Text Classification 20NEWS RMDL Accuracy 91.21 # 1
Image Classification CIFAR-10 RMDL Percentage correct 91.21 # 42
Image Classification CIFAR-10 RMDL Percentage error 8.79 # 18
Document Classification IMDb RMDL Accuracy 90.79 # 1
Image Classification MNIST RMDL Percentage error 0.18 # 1
Face Recognition Olivetti Faces 5 Image RMDL Accuracy 95.00 # 1
Document Classification Reuters-21578 RMDL Accuracy 90.69 # 1
Document Classification WOS-11967 RMDL Accuracy 91.59 # 1
Document Classification WOS-46985 RMDL Accuracy 90.69 # 1
Document Classification WOS-5736 RMDL Accuracy 93.57 # 1