https://github.com/thefloatingstring/ailink
AILink machine learning API
https://github.com/thefloatingstring/ailink
Last synced: 11 months ago
JSON representation
AILink machine learning API
- Host: GitHub
- URL: https://github.com/thefloatingstring/ailink
- Owner: TheFloatingString
- Created: 2019-08-16T05:22:27.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-08-16T13:25:59.000Z (over 6 years ago)
- Last Synced: 2025-01-13T08:12:47.993Z (about 1 year ago)
- Language: Python
- Homepage:
- Size: 1.68 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AILink
API using a 1D Convolutional Neural Network to provide sentimental analysis based on the Kaggle Rotten Tomatoes Movie Review Challenge (https://www.kaggle.com/c/sentiment-analysis-on-movie-reviews), with an accuracy of ~60% with word count vectorial encoding.
## Quickstart
Go to `http://ailinkdev.herokuapp.com/compute/this is a sample text`
## Running Flask API
```bash
pip3 install requirements.txt
python3 app.py
```
Then, enter in browser the following URL:
`http://localhost:5000/compute/`
which will return a JSON of probabilities of classes 0 (terrible) to 4 (excellent)
`{"0":0.1369616836309433,"1":0.2723471224308014,"2":0.4098854064941406,"3":0.0,"4":0.0}`
## Running Machine Learning Model
```bash
pip3 install requirements.txt
python3 src/train_cnn.py
```
## Running data visualization
```bash
pip3 install requirements.txt
python3 src/plot_data.py
```