2018 International Conference on Artificial Intelligence and Data Processing (IDAP) 2018
DOI: 10.1109/idap.2018.8620869
|View full text |Cite
|
Sign up to set email alerts
|

Experiments on Fine Tuning Deep Learning Models With News Data For Tweet Classification

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1

Citation Types

0
2
0
1

Year Published

2020
2020
2021
2021

Publication Types

Select...
2
2
1

Relationship

1
4

Authors

Journals

citations
Cited by 5 publications
(3 citation statements)
references
References 5 publications
0
2
0
1
Order By: Relevance
“…Her kategoriden 100.000 adet olmak üzere Ekonomi, Spor, Siyaset ve Kültür haberlerinden oluşan toplam 400.000 doküman bulunmaktadır. Bu veri seti Bigailab-5news-500K'nın bir alt kümesidir [12].…”
Section: Veri Seti Yöntem Ve Hesaplama Ortamıunclassified
“…Her kategoriden 100.000 adet olmak üzere Ekonomi, Spor, Siyaset ve Kültür haberlerinden oluşan toplam 400.000 doküman bulunmaktadır. Bu veri seti Bigailab-5news-500K'nın bir alt kümesidir [12].…”
Section: Veri Seti Yöntem Ve Hesaplama Ortamıunclassified
“…While there is no rule that defines which model is the right one for a specific task, the task objective and the input characteristics play an important role in narrowing the set of models candidates. For example, the RNN has two limitations which are exploding gradient and vanishing gradient [24]. The vanishing gradient could accrue with a long sequence input, which is the case in document classification.…”
Section: The Deep Learning Approachmentioning
confidence: 99%
“…Our model, shown in Figure 4, consists of four layers: an embedding layer, a 1D CNN layer, an LSTM layer, and a softmax layer. The model starts with feeding the document as raw input features into the embedding layer, which passes its embeddings to the CNN layer; the CNN layer derives important features and passes them to the LSTM layer [24]. Finally, the output of the LSTM layer is fed to the softmax layer that determines the class of the input document.…”
Section: The Deep Learning Approachmentioning
confidence: 99%