Abstract:
Movie plot summaries are expected to reflect the genre of movies since many spectators read the plot summaries before deciding to watch a movie. In this study, we perform...Show MoreMetadata
Abstract:
Movie plot summaries are expected to reflect the genre of movies since many spectators read the plot summaries before deciding to watch a movie. In this study, we perform movie genre classification from plot summaries of movies using bidirectional LSTM (Bi-LSTM). We first divide each plot summary of a movie into sentences and assign the genre of corresponding movie to each sentence. Next, using the word representations of sentences, we train Bi-LSTM networks. We estimate the genres for each sentence separately. Since plot summaries generally contain multiple sentences, we use majority voting for the final decision by considering the posterior probabilities of genres assigned to sentences. Our results reflect that, training Bi-LSTM network after dividing the plot summaries into their sentences and fusing the predictions for individual sentences outperform training the network with the whole plot summaries with the limited amount of data. Moreover, employing Bi-LSTM performs better compared to basic Recurrent Neural Networks (RNNs) and Logistic Regression (LR) as a baseline.
Date of Conference: 31 January 2018 - 02 February 2018
Date Added to IEEE Xplore: 12 April 2018
ISBN Information:
Keywords assist with retrieval of results and provide a means to discovering other relevant content. Learn more.
- IEEE Keywords
- Index Terms
- Bidirectional Long Short-term Memory ,
- Summary Plots ,
- Movie Genres ,
- Genre Classification ,
- Neural Network ,
- Recurrent Neural Network ,
- Majority Voting ,
- Word Representations ,
- Limited Amount Of Data ,
- Bidirectional Long Short-term Memory Network ,
- Training Dataset ,
- Convolutional Neural Network ,
- Classification Task ,
- Visual Features ,
- Multi-label ,
- Class Labels ,
- Stochastic Gradient Descent ,
- Word Embedding ,
- Softmax Layer ,
- Textual Features ,
- Term Frequency-inverse Document Frequency ,
- Syntactic Relations ,
- Recurrent Neural Network Model ,
- Number Of Sentences ,
- F-score Values ,
- Full Plots
- Author Keywords
Keywords assist with retrieval of results and provide a means to discovering other relevant content. Learn more.
- IEEE Keywords
- Index Terms
- Bidirectional Long Short-term Memory ,
- Summary Plots ,
- Movie Genres ,
- Genre Classification ,
- Neural Network ,
- Recurrent Neural Network ,
- Majority Voting ,
- Word Representations ,
- Limited Amount Of Data ,
- Bidirectional Long Short-term Memory Network ,
- Training Dataset ,
- Convolutional Neural Network ,
- Classification Task ,
- Visual Features ,
- Multi-label ,
- Class Labels ,
- Stochastic Gradient Descent ,
- Word Embedding ,
- Softmax Layer ,
- Textual Features ,
- Term Frequency-inverse Document Frequency ,
- Syntactic Relations ,
- Recurrent Neural Network Model ,
- Number Of Sentences ,
- F-score Values ,
- Full Plots
- Author Keywords