Stock Sentiment Analysis using News Headlines

Data Preparation

• Loaded the dataset and split it into training and testing sets.
• Removed punctuations and converted headlines to lowercase.

Feature Extraction

• Used CountVectorizer to implement the Bag of Words model with bigrams.

Model Training

• Trained a RandomForest Classifier with 200 estimators using the training data.

Model Evaluation

Confusion Matrix: [[138 48] [ 7 185]] Accuracy Score: 0.8544973544973545 Classification Report: precision recall f1-score support

       0       0.95      0.74      0.83       186
       1       0.79      0.96      0.87       192

accuracy                           0.85       378

macro avg 0.87 0.85 0.85 378 weighted avg 0.87 0.85 0.85 378

Buy/Sell Signal Generation

• Generated buy and sell signals based on the sentiment predictions using moving crossover startegy.

Portfolio Returns and Performance Metrics

Moving Average Strategy: Sharpe Ratio: 1.585533826042266 Maximum Drawdown: -0.2849368491361862 Number of Trades Executed: 17 Win Ratio: 0.23529411764705882

Plot

• Plotted the stock price with buy and sell signals
• 

portfolio returns