Welcome to my website!
I am a junior at The Quarry Lane School, pursuing a STEM curriculum.
QuikSum is a summarizer for English-language texts. It takes as input an arbitrary block of text, say, an essay or an article (of any length) and identifies key sentences that capture the essence of what the author is trying to convey.
At its core, the program assigns a score to each sentence in the text, which is derived from the scores of the words in the sentence. The summary generated as output is then the set of sentences with the highest scores. The length of the summary is controlled via another input to the program.
A lookup table of pre-computed word scores is used to calculate the sentence scores. The table of word scores is based on the frequency and novelty of each word, i.e., common words like "the", "him", "she", etc. are assigned low scores, and words that occur less frequently in prose, like "politics", "domestic", "scrimmage", etc. are assigned higher scores. These scores are based on the frequency of occurrence of each word on reputable public websites like news outlets and official sources of information.
The core logic for the above algorithm was implemented in Swift. An iOS app was developed in the simulator to expose this functionality.