Logo Lanfrica

Temesgenzewude/Amharic-Text-Summarizer-Using-Machine-Learning

Domaine:

natural language processing

Type de record:

software
Créateur:
Tem
Hôte:
# Amharic text summarizer Team Members Aklile Yilma UGR/7107/12 Joshua Tesfaye UGR/0359/12 Temesgen Zewude UGR/3848/12 Abinet Anamo UGR/7110/12 Henok Mekuanint UGR/2272/12 Algorithm 1: Extraction 1. Extract all the sentences from text. 2. Extract all the words from text. 3. Assign a score to each word. 4. Assign a score to each sentence. 5. Put the sentences with the highest score together in chronological order to produce the summary. Algorithm 2: Cosine Similarity 1. TF-IDF weights to each individual word in a sentence 2. Generate cosine-similarity of each TF-IDF sentence pair matrix 3. Average the weights of each vector 4. Vectors with highest average summarize the text words with higher weights (more unique) have more importance Installation >pip install -r requirements.txt >python main.py pdf at: amnewsupdate.files.wordpres… Reference links: - towardsdatascience.com - github.com - blog.floydhub.com - machinelearningplus.com