Goal:
Create a product to highlight a prediction algorithm and to provide an interface that can be accessed by others.
Description of the algorithm: Based on the data provided by SwiftKey we have done the following.
- Sampling the data. We do not need the entire dataset because that would increase the time of response of the model.
- Clean the data and filter bad words.
- Split the data in train and test sets to train our algorithm.
- Create n-grams (the max was quadgram) to check which combinations of words are the most likely to appear in a phrase.
- Deploy the algorithm in an app.
