Poems-Genre-Predictor-NLP-

Natural Language Processing: Predicting Genres of Poems

Poems Classifier

Data Source: https://www.kaggle.com/ultrajack/modern-renaissance-poetry

The dataset consists of the following data:

1. Total no. of data points: 573
2. Number of Attributes/Columns in data: 5
3. Number of Authors: 67
4. Number of different poems: 506
5. Number of different poem names: 509
6. Types(genres): 3

Attribute Information:

1. author
2. content: Poems
3. poem Name
4. age: Modern and Renaissance
5. type: Genre(Mythology & Folklore, Nature and Love)

Loading the data: The dataset is available in .csv file.

Approach: In steps:-

1. Removing the duplicate and Null Data.
2. Each poem is converted to lower case.
3. Removal of punctuations like --->,/./‘/: etc
4. Removal of unimportant words like “A”,"An","The","Aboard","About","Above","Absent","Across","After", etc
5. Vectorizing using TFIDF Vectorizer.
6. Splitting the data into the 80/20 ratio.
7. The first classifier used-----> XGBoost
8. Second Classification using SVM

Packages used:

1. Numpy
2. Pandas
3. Matplotlib
4. Seaborn
5. Regular expression--->re
6. xgboost From Scikit-Learn:
7. TfidfVectorizer
8. train_test_split
9. LabelEncoder
10. confusion_matrix
11. accuracy_score
12. LinearSVC
13. average_precision_score