This repository contains end-to-end code to automatically identify sentiment on a under-resource language (Roman-Urdu) on social media using machine learning.
A data corpus comprising of more than 20000 records in Roman Udu (a limited resource language) was collected and tagged for Sentiment (Positive, Negative, Neutral).
Link to data: http://archive.ics.uci.edu/ml/datasets/Roman+Urdu+Data+Set
├── LICENSE
├── README.md <- The top-level README for developers using this project.
├── data
│ ├── external <- Data from third party sources.
│ ├── interim <- Intermediate data that has been transformed.
│ ├── processed <- The final, canonical data sets for modeling.
│ └── raw <- The original, immutable data dump.
│
├── docs <- A default Sphinx project; see sphinx-doc.org for details
│
├── models <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks <- Jupyter notebooks.
│ ├── 1.0-cchen-exploratory-data-analysis
│ ├── 2.0_cchen_roman_urdu_ngram_models
│
├── references <- Data dictionaries, manuals, and all other explanatory materials.
│
├── reports <- Generated analysis as HTML, PDF, LaTeX, etc.
│ └── figures <- Generated graphics and figures to be used in reporting
│
├── requirements.txt <- The requirements file for reproducing the analysis environment, e.g.
│ generated with `pip freeze > requirements.txt`
│
├── src <- Source code for use in this project.
│ ├── __init__.py <- Makes src a Python module
│ │
│ ├── data <- Functions to load data from roman urdo sentiment analysis dataset
│ │ └── make_dataset.py
│ │
│ ├── features <- Functions to perform N-gram and sequence vectorization functions
│ │ └── vectorize_data.py
│ │
│ ├── models <- Functions to build, train, and tune models with different hyperparameters
│ │ │
│ │ ├── build_model.py
│ │ ├── train_mlp_model.py
│ │ ├── train_sequence_model.py
│ │ └── tune_mlp_model.py
│ │
│ └── visualization <- Functions to create exploratory and results oriented visualizations
│ └── visualize.py