Introduction to Text Prediction App
Charlie Zuo
April 25th, 2015
Summary
- Data Processing and Tokenization
- Raw Text Data was converted to dataset for analysis
- N-gram Modelling using Markov Chain and Smoothing technique
- Application was deloyed as a web application
Predicting Algorithm
Markov Chain
- We use Markov Chain in our n-gram model. This process is used to predict the next word when given current words.
Katz Back-off model
- A generative ngram language model that estimates the conditional probability of a word given its history.
Instructions for this app
- Simply type your words/sentence into the text field of app.
System will truncate the words you just typed
and use them as input for the predictive algorithm.
- As shown in this image
Where to get it
- Technology used in this app:
- You can use the app from the link blow
- Text Prediction
- (Note: It may take a few seconds for app to load, please be patient. It has been tested, and working.) Enjoy!