Google SPS 2020 Portfolio

Join a faction to fight deadlines and the dead!

vanilla implementation of Cocke–Kasami-Younger parser

convert context free grammar in NLTK format to chomsky normal form

implementation of improved probabilistic Cocke-Kasami-Younger parser and grammar induction script with parent annotation optimisation

Implementation of a multi-variate Bernoulli Naive Bayes model and a multi-nominal Naive Bayes model for classification

build a decision tree classifier with max depth and minimum information gain passed in as early stop criteria

Implementation of a kNN classifier with two different distance metric options: Euclidean distance and Cosine distance

German compound word morphological splitting. Reproducing results from Koehn and Knight (2003)

Implementing an English POS tagger using HMM and Viterbi

Build a language model for English from ngrams

Reproduce morphological segmentation inside-out results from Cotterell et al. (2016), and a frequency based metric using the monolingual German corpus in described in Koehn and Knight (2003).

Build a Hidden Markov Model from different ngrams and interpolation smoothing

Vanilla sentence boundary detection using regex and an exception list. Evaluation with f1 score

Vanilla English tokeniser using regex and an exception list

Scrape movie title's different madarin translations in Taiwan, China and Hong Kong

scrape data from facebook news pages without using graph api

Hack it till you make it. A static site with notes on data science, python, NLTK.