Amharic semantic-based IR using BM25 was developed
Best insight from top research papers
Yes, an Amharic semantic-based IR system using BM25 was developed. The system utilizes query expansion based on a deep neural learning model and WordNet to optimize retrieval results. The development of the system involved the construction of resources and corpora for evaluation, as well as the use of root-based text representation for improved performance. Comparisons with stem-based representation showed that using root forms of words is better for both resource construction and system development. The effectiveness of the system was evaluated on the Amharic Adhoc Information Retrieval Test Collection (2AIRTC) .
Answers from top 4 papers
More filters
| Papers (4) | Insight |
|---|---|
01 May 2017 7 Citations | The provided paper does not mention the development of an Amharic semantic-based IR using BM25. |
| The provided paper does not mention the development of an Amharic semantic-based IR using BM25. | |
| The paper does not mention the development of an Amharic semantic-based IR using BM25. | |
| The paper does not mention the development of an Amharic semantic-based IR system using BM25. |
Related Questions
Amharic Speech Emotion Recognition using deep learning4 answersAmharic Speech Emotion Recognition (SER) can be effectively achieved through deep learning techniques. Research has shown that deep learning algorithms like Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and a combination of CNN and BiLSTM are particularly beneficial for sentiment analysis in the Amharic language. These algorithms excel in extracting features and processing sequential information, leading to high accuracy rates in sentiment classification. Additionally, the use of deep learning models, such as CNN and BiLSTM, has proven successful in real-time speech emotion recognition tasks, showcasing the effectiveness of these approaches in identifying emotional states from speech signals. Leveraging deep learning techniques tailored for Amharic sentiment analysis can significantly enhance the accuracy and efficiency of speech emotion recognition systems.
Amharic Speech Recognition using deep learning4 answersAmharic speech recognition using deep learning has shown promising results in various studies. One research focused on sentiment analysis in Amharic text, utilizing deep learning algorithms like CNN and BiLSTM, achieving an accuracy rate of 91.60%. Another study addressed the challenge of recognizing handwritten and machine-printed characters in Amharic by proposing a CNN model, achieving 96.02% accuracy. These findings highlight the effectiveness of deep learning techniques in processing and understanding the complexities of the Amharic language. While specific research on speech recognition in Amharic is limited, the success of deep learning models in related tasks suggests a strong potential for Amharic speech recognition systems to be developed using similar approaches.
What are the NLP techniques commonly used in Amharic compared to those in large language models for English?4 answersNLP techniques commonly used in Amharic include word embeddings and BERT language models. Word embeddings such as word2vec, GloVe, and fastText have been explored for learning text representations in Amharic. These techniques have been used for query expansion and language modeling tasks in Amharic information retrieval and NLP. Additionally, a pre-trained Amharic BERT model has been utilized for masked language modeling, next sentence prediction, and text classification tasks in Amharic. Experimental results show that word-based representations and fastText perform well in Amharic text processing.
Amharic Text news classification5 answersAmharic text classification has been studied in the context of text complexity and machine learning algorithms. A complexity classification model for Amharic texts was developed using supervised machine learning, achieving an accuracy of 87.1% with Support Vector Machine (SVM) algorithm. Arabic text classification methods have also been extensively studied, given the vast volume of Arabic textual material available on the internet. Empirical studies have been conducted on Arabic text classification using various classification models and feature vectorization methods, with SVM and Logistic Regression (LR) showing the highest performance. Additionally, deep learning models such as Multilayer Perceptron (MLP) and Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) have been used for Arabic multi-label text classification, achieving F1 scores of 82.03 and 80.37 respectively.
Amharic text news classification5 answersText classification for Amharic news documents has been addressed in several papers. One paper proposed a hybrid feature selection method using document frequency (DF) and genetic algorithm (GA) for Amharic text classification. This method outperformed other feature selection methods and improved classification accuracy when combined with the Extra Tree Classifier (ETC). Another paper introduced a news text classification method based on the GRU_CNN model, which combines the advantages of CNN and GRU. This hybrid model outperformed single CNN, LSTM, and GRU models in terms of classification effect and accuracy. These papers provide valuable insights and techniques for effectively classifying Amharic text news documents.
Amharic Hate Speech Detection4 answersAmharic hate speech detection has been addressed in multiple studies. One study applies a multi-modal approach using deep learning techniques to detect hate speech in Amharic on social media. The study collects audio data from YouTube videos, transcribes them into text using the Google Speech-to-Text API, and extracts textual and acoustic features for analysis. Four deep learning algorithms are used, with the multi-modal model employing BILSTM achieving the highest accuracy of 88.15%. Another study focuses on creating a Tunisian reference dataset for hate speech and abuse detection, as the Tunisian dialect is underrepresented. The study presents a neural network model based on AraBERT, which achieves a high F1 score of 0.99. These studies demonstrate efforts to detect and address hate speech in Amharic and Tunisian languages on social media.