Meter classification of Arabic poems using deep bidirectional recurrent neural networks

Al-shaibani, Maged S.; Alyafeai, Zaid; Ahmad, Irfan

doi:10.1016/j.patrec.2020.05.028

Cited by 23 publications

(12 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Yousefi [ 29 ] finds the unwritten linking vowel (izafe) by convolutional neural networks. Al-shaibani et al [ 30 ] by deep bidirectional recurrent neural networks classify the meter of Arabic poems without diacritizing. Abandah et al [ 31 ] use recurrent neural networks with bidirectional long short-term memory cells for diacritizing the input Arabic poems.…”

Section: Background and Related Workmentioning

confidence: 99%

Automatic meter classification of Kurdish poems

Mahmudi

Veisi

2023

PLoS ONE

View full text Add to dashboard Cite

Most of the classic texts in Kurdish literature are poems. Knowing the meter of the poems is helpful for correct reading, a better understanding of the meaning, and avoiding ambiguity. This paper presents a rule-based method for the automatic classification of the poem meter for the Central Kurdish language also known as Sorani. The metrical system of Kurdish poetry is divided into three classes quantitative, syllabic, and free verses. As the vowel length is not phonemic in the language, there are uncertainties in syllable weight and meter identification. The proposed method generates all the possible situations and then, by considering all lines of the input poem and the common meter patterns of Kurdish poetry, identifies the most probable meter type and pattern of the input poem. Evaluation of the method on a dataset from VejinBooks Kurdish corpus resulted in 97.3% of precision in meter type and 96.2% of precision in pattern identification.

show abstract

Section: Background and Related Workmentioning

confidence: 99%

Automatic meter classification of Kurdish poems

Mahmudi

Veisi

2023

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Sudanese vocabulary is mostly inspired by MSA, but with important Greek, Turkish and English modifications to the phonology. The morphology of Sudanese words shares many features with MSA, but the method of dialect inflection is Saudi (3C) 90+10 CNN 86.54% [43] Moroccan (2C) 90+10 Majority Voting 83.45% [23] Egyptian, Iraqi and Levantine (3C) 80+10+10 LSTM 71.4 % [6] Jordanian (2C) 90+10 Ensemble 93.4% [5] Lebanon (2C) 80+20 LR 89.80% [1] Algerian (2C) 85+15 SVM 0.86% [31] Tunisian (2C) 80+10+10 Deep-LSTM 90.00% [31] JEG, TAC and TSAC (2C) 90+10 Tw-StAR 82.08% [25] Egyptian, MSA (2C)(n-C) 80+10+10 MC1, MC2 92.96% [7] Modern Standard Arabic 85+15 BiGRU 94.32% [3] Modern Standard Arabic 80+20 RF+SMOTE 96.00% more complicated than MSA in some respects [20]. Following a thorough study of such dialect differences, we have created two datasets based on social media posts, built a CNN-based model for sentiment analyis, and applied it to the datasets.…”

Section: Introductionmentioning

confidence: 99%

A Deep CNN Architecture with Novel Pooling Layer Applied to Two Sudanese Arabic Sentiment Datasets

Mustafa¹,

Sutcliffe²,

Sun³

et al. 2022

Preprint

View full text Add to dashboard Cite

Arabic sentiment analysis has become an important research field in recent years. Initially, work focused on Modern Standard Arabic (MSA), which is the most widely-used form. Since then, work has been carried out on several different dialects, including Egyptian, Levantine and Moroccan. Moreover, a number of datasets have been created to support such work. However, up until now, less work has been carried out on Sudanese Arabic, a dialect which has 32 million speakers. In this paper, two new publicly available datasets are introduced, the 2-Class Sudanese Sentiment Dataset (SudSenti2) and the 3-Class Sudanese Sentiment Dataset (SudSenti3). Furthermore, a CNN architecture, SCM, is proposed, comprising five CNN layers together with a novel pooling layer, MMA, to extract the best features. This SCM+MMA model is applied to SudSenti2 and SudSenti3 with accuracies of 92.75% and 84.39%. Next, the model is compared to other deep learning classifiers and shown to be superior on these new datasets. Finally, the proposed model is applied to the existing Saudi Sentiment Dataset and to the MSA Hotel Arabic Review Dataset with accuracies 85.55% and 90.01%.

show abstract

“…It varies greatly from language to language, which implies that algorithms designed for poetry analysis in one language more often may not be equally applicable to another. In contrast to Hindi, a significant amount of work has been done in computing poetic elements of other languages, such as metre detection and classification of Arabic [1,4,18] and Persian poetry [28], a study of metre as a stylistic feature in Latin poetry [6], an expert system for harmony test of Arabic poetry [3], a statistical evaluation of Chinese Tang [2,32] and English [10] poetry, an emotion based classification for Marathi [26], Punjabi [8], and Arabic [20] poetry, a study of rhythm of Tibetan poetry [14]. One of the surprises of this study is that none of the above languages has a computational system or tool that recognises and quantifies figure of speech, a very important component of poetry, which is one of the novelties of this article.…”

Section: Introductionmentioning

confidence: 99%

Design and Development of Computational Tools for Analyzing Elements of Hindi Poetry

Naaz

Singh

2022

IEEE Access

View full text Add to dashboard Cite

Poetry writing is a qualitative subject and so is its analysis. Mapping of these poetic elements onto a scale of real number is a lacking necessity. Albeit, Hindi literary heritage, being so huge and glorified, there is remarkably very few computational works done exploring the underlying structures. Out of which most of them is to detect a particular metre rather than a generalized approach. The state-of-art metadata generator fails to provide any measures of underlying structural elements of poetry. There is no automated system that generates rhyming pattern hidden in a poem for Hindi language or a system to detect and estimate the extent of figure of speech in a given text of any language. In this article, to extract and evaluate elements of poetry, three efficient tools, namely Text2Mātrā, RPaGen and FoSCal, have been designed and developed. The Text2Mātrā tool provides the numeral scansion for any Hindi input text, which can serve as basis for copious analytical and detection work. RPaGen detects the poem type of any input poem and outputs its rhyming pattern. FoSCal gives a quantitative representation of detected figures of speech in any input text, using the scoring scheme formulated using fuzzy approach and weighted analysis. These tools may find their utility in various fields such as education, literary criticism, philology, authorship-attribution, etc. There have been various computational activities done in the field of poetry analysis over the various languages across the world. However, quantifying the extent of Figure of Speech in poetic compositions, in any language, is entirely a novel approach. Mapping the aesthetic properties of a subjective idea (like poetry) onto a numeral scale, to the best of our knowledge, is first of its kind for Hindi language.

show abstract

Meter classification of Arabic poems using deep bidirectional recurrent neural networks

Cited by 23 publications

References 7 publications

Automatic meter classification of Kurdish poems

Automatic meter classification of Kurdish poems

A Deep CNN Architecture with Novel Pooling Layer Applied to Two Sudanese Arabic Sentiment Datasets

Design and Development of Computational Tools for Analyzing Elements of Hindi Poetry

Contact Info

Product

Resources

About