An Application of Ensemble and Deep Learning Models in Predictive Analytics

Sonbhadra, Sanjay Kumar; Agarwal, Sonali; Syafrullah, Mohammad; Adiyarta, Krisna

doi:10.1109/iciea49774.2020.9102115

Cited by 3 publications

(1 citation statement)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More variance shows over-fitting [4], hence in an ideal model less bias and less variance is expected. Seeking for these objectives, the ensemble methods are established as the best choice for image classification [30]. Therefore, in the present article an ensemble technique is proposed for fruit classification.…”

Section: Introductionmentioning

confidence: 99%

Fruit classification using deep feature maps in the presence of deceptive similar classes

Dandekar,

Punn,

Sonbhadra

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Autonomous detection and classification of objects are admired area of research in many industrial applications. Though, humans can distinguish objects with high multi-granular similarities very easily; but for the machines, it is a very challenging task. The convolution neural networks (CNN) have illustrated efficient performance in multi-level representations of objects for classification. Conventionally, the existing deep learning models utilize the transformed features generated by the rearmost layer for training and testing. However, it is evident that this does not work well with multi-granular data, especially, in presence of deceptive similar classes (almost similar but different classes). The objective of the present research is to address the challenge of classification of deceptively similar multi-granular objects with an ensemble approach thfat utilizes activations from multiple layers of CNN (deep features). These multi-layer activations are further utilized to build multiple deep decision trees (known as Random forest) for classification of objects with similar appearance. The Fruits-360 dataset is utilized for evaluation of the proposed approach. With extensive trials it was observed that the proposed model outperformed over the conventional deep learning approaches.

show abstract

Section: Introductionmentioning

confidence: 99%