2. Overview of what we have done?
1. Scraped poems of various authors from https://mypoeticside.com/poets/
2. Preprocessed those poems, cleaned them and created a dataframe of all
the lines.
3. Using spacy(open-source NLP library) and lstm(RNN) we have generated
poems from a theme given by user.
4. Using flask to create a basic web-app as frontend.
6. Poem Generation
Using Spacy
To find similarity between
our theme word and poems
we use a large corpus of
words included in spacy.
“en_core_web_lg”
7. Generation Algorithm
Select a initial word
First we randomly select n verses
Then we choose the verse with highest
similarity to initial word
Compute similarity between initial words and
all n verses
Set initial word
to that verse
11. Poem Generation
Using LSTM(tried)
LSTM (Long Short-Term Memory)
network is a type of RNN (Recurrent
Neural Network) that is widely used
for learning sequential data
prediction problems. As every other
neural network LSTM also has some
layers which help it to learn and
recognize the pattern for better
performance.