• CyberBullying8020.ipynb : 0.7342498450804434/ 0.7342329936656569
• CyberBullying7525.ipynb : 0.7342587152369761/ 0.7342097532314924
• CyberBullying7030.ipynb : 0.7342356520826776/ 0.7342717258261934
• CyberBullying6535.ipynb : 0.7342447953447643/ 0.7342495934532864
• CyberBullying6040.ipynb : 0.7342554623905991/ 0.7342329936656569

• CyberBullying8020.ipynb
• CyberBullying7525.ipynb
• CyberBullying7030.ipynb
• CyberBullying6535.ipynb
• CyberBullying6040.ipynb

• CyberBullying8020.ipynb : 0.7342498450804434/ 0.7342329936656569
• CyberBullying7525.ipynb : 0.7342831962397179/ 0.7342097532314924
• CyberBullying7030.ipynb : 0.7342356520826776/ 0.7342717258261934
• CyberBullying6535.ipynb : 0.7342447953447643/ 0.7342495934532864
• CyberBullying6040.ipynb : 0.7342554623905991/ 0.7342329936656569

CyberBullying8020CyberTrollIEEEDataset.ipynb

Length: 22141
Dataset:

1. cybertroll_dataset.csv
2. CyberBullyingTypesDataset.csv

55K (54464)

CyberBullying8020.ipynb : 0.8591723853021506/ 0.8122647571835123
CyberBullying7525.ipynb : 0.8608499804151978/ 0.8116186839012925
CyberBullying7030.ipynb : 0.8621078585667821/ 0.8096695226438189
CyberBullying6535.ipynb : 0.864495353238609/ 0.808372239416671
CyberBullying6040.ipynb : 0.8648020074667973/ 0.808409070044983

Final Dataset:

• cyberbullying_tweets.csv
• CyberBullyingTypesDataset.csv
• cybertroll_dataset.csv
• classified_tweets.csv

• CyberBullying8020.ipynb : 0.7342498450804434/ 0.7342329936656569
• CyberBullying7525.ipynb : 0.7342831962397179/ 0.7342097532314924
• CyberBullying7030.ipynb : 0.7342356520826776/ 0.7342717258261934
• CyberBullying6535.ipynb : 0.7342447953447643/ 0.7342495934532864
• CyberBullying6040.ipynb : 0.7342554623905991/ 0.7342329936656569

Length: 81456

Dataset:

1. cybertroll_dataset.csv
2. CyberBullyingTypesDataset.csv
3. cyberbullying_tweets.csv
4. classified_tweets.csv

Graph of cybertroll_dataset.csv & CyberBullyingTypesDataset.csv & cyberbullying_tweets.csv & classified_tweets.csv

WordCloud

• CyberBullying8020.ipynb : 0.9446191274012531/ 0.7636096575782613
• CyberBullying7525.ipynb : 0.9481737171954563/ 0.7660105757931844
• CyberBullying7030.ipynb : 0.9511856048683244/ 0.7678702570379436
• CyberBullying6535.ipynb : 0.9540408463037767/ 0.7668782458165032
• CyberBullying6040.ipynb : 0.9573107289307791/ 0.7656293032222529

Creating the final ML Model for Cyberbullying Detection

CBDA: Web App for Cyberbullying Detection using NLP & ML

• CyberBullying8020.ipynb : 0.7987652337563976/ 0.7805930414027357
• CyberBullying7525.ipynb : 0.7966118292205249/ 0.7791568742655699
• CyberBullying7030.ipynb : 0.7964274472773056/ 0.7768665850673194
• CyberBullying6535.ipynb : 0.7949775430072596/ 0.7746944342443477
• CyberBullying6040.ipynb : 0.7942346532835547/ 0.7732029743872212

k-fold cross-validation helps to evaluate the performance of different algorithms.

CyberBullying8020CyberTrollDataset.ipynb

Length: 20001
Dataset:

1. cybertroll_dataset.csv

Grid Search helps to find the best combination of hyperparameters for each algorithm.

Best Accuracy per Algorithm in any Dataset

CyberBullying8020IEEEDataset.ipynb

Length: 2140
Dataset:

1. CyberBullyingTypesDataset.csv