added images and updated readme

Ignatiocalvin · Jun 11, 2024 · 3333262 · 3333262
1 parent 5711a52
commit 3333262
Show file tree

Hide file tree

Showing 16 changed files with 68 additions and 24 deletions.
diff --git a/README.md b/README.md
@@ -1,14 +1,39 @@
 # LLM_StockPredictor
 ## Overview
+The impact of Natural Language Processing (NLP) algorithms in predicting
+stock market prices, especially price shocks.
+
+Commodity price shocks are times when the prices for commodities have
+drastically increased or decreased over a short time. Typically, the stock market
+and economic performance are aligned. Thus, when the stock market is performing well, it is usually a function of a growing economy.
+
+Stock market declines have the potential to diminish wealth across personal
+and retirement investment portfolios. Consequently, individuals witnessing a depreciation in their portfolio value are inclined to curtail their expenditures. 
+
+With this project, we aim to develop a model that analyzes news headlines
+and predicts stock market crashes based on text data. Having such a model may
+help us anticipate the stock market movement to better manage our wealth and
+prepare for adverse economic events. We chose to work with NLP algorithms and
+fine-tune existing pre-trained LLMs, as classical autoregressive models show
+poor predictive capacity during price shocks. Using textual data as input such
+as news headlines may help the model adjust its predictions to keep up with
+drastically changing trends. Aside from that, we used also analyzed the effects of sentiment analysis and moving averages as features used for prediction. The models are evaluated based on the root mean squared error (RMSE) of the predicted stock prices. 
+
+## Data
+
+The data is obtained from the Kaggle dataset [Daily News for Stock Market Prediction](https://www.kaggle.com/aaron7sun/stocknews). The dataset contains historical news headlines from Reddit WorldNews Channel and Dow Jones Industrial Average (DJIA) stock prices. The news headlines are from 2008 to 2016, and the stock prices are from 2008 to 2016. Combined_News_DJIA.csv contains the top 25 news headlines and the corresponding stock prices for each day. upload_DJIA_table.csv contains the stock prices for each day. 
+
+The project is divided into the following sections:
 
 ## Notebooks
-### 1. Univariate Model
-Filename: `univariate.ipynb`
-Description: This notebook focuses on using a Long Short-Term Memory (LSTM) neural network to predict stock prices based on a univariate time series.
 
-### 2. FinGPT Sentiment Retrieval
+### 1. FinGPT Sentiment Retrieval
 Filename: `retrieve_sentiment.ipynb`
-Description: This notebook demonstrates the use of a Large Language Model (LLM), specifically FinGPT, to retrieve and analyze sentiments of financial news articles using the LangChain framework. We asked FinGPT to generate a sentiment score for each news headline in the dataset. The sentiment score ranges from -10 to 10, where -1 indicates a negative sentiment, 0 indicates a neutral sentiment, and 1 indicates a positive sentiment.
+Description: This notebook demonstrates the use of a LLM, specifically FinGPT, to retrieve and analyze sentiments of financial news articles using the LangChain framework. We asked FinGPT to generate a sentiment score for each news headline in the dataset. The sentiment score ranges from -10 to 10, where -1 indicates a negative sentiment, 0 indicates a neutral sentiment, and 1 indicates a positive sentiment.
+
+### 2. Univariate Model
+Filename: `univariate.ipynb`
+Description: This notebook focuses on using a LSTM neural network to predict stock prices based on a univariate time series.
 
 ### 3. Predicting Stock Prices with News Headlines without Looking Back at the Data
 Filename: `news_rnn.ipynb`
@@ -28,24 +53,51 @@ Description: Here, aside from using the look back window of 50 days of news head
 Filename: `bert_MA.ipynb`
 Description: In this notebook, we predict stock prices using an LSTM neural network and financial news headlines which uses BERT embeddings. We use a look back window of 50 days of news headlines and stock prices to predict the next day's stock price in the training data and a 30 day window in the testing data. We also include a moving average of the stock prices of 10 days as an additional feature in the model.
 
+### 7. Predicting Stock Prices with FastText Embeddings
+Filename: `lookback_fasttext.ipynb`
+Description: This notebook predicts stock prices using an LSTM neural network and financial news headlines which uses FastText embeddings. We use a look back window of 50 days of news headlines and stock prices to predict the next day's stock price in the training data and a 30 day window in the testing data. The FastText embeddings are used as an alternative to BERT embeddings.
+
+### 8. Predicting Stock Prices with XGBoost
+Filename: `xgboost.ipynb`
+Description: This notebook predicts stock prices using an XGBoost model. The model uses the stock prices and news headlines as features to predict the stock prices. The model is trained on the training data and evaluated on the testing data.
+
+
+
+
+
 ## Results
-For our models, we used the root mean squared error (RMSE) as the evaluation metric. The RMSE is a measure of the differences between the predicted values and the actual values. It gives us an idea of how well the model is performing in terms of predicting the stock prices.
+For our models, we used the RMSE as the evaluation metric. The RMSE is a measure of the differences between the predicted values and the actual values. It gives us an idea of how well the model is performing in terms of predicting the stock prices.
 
 The table below shows the RMSE values for the training and testing data for each of the models.
 
 <center>
 
-| Notebooks 	        | train RMSE 	        | test RMSE 	        |
+| Notebooks 	        | Train RMSE 	        | Test RMSE 	        |
 |:---:	                |:---:	                |:---:	                |
-| bert_MA 	            | 10.83293233874024 	| 816.5899384610334 	|
-| lookback_news_rnn 	| 15.457514085649745 	| 309.5794180019174 	|
 | univariate 	        | 31.743497936455558 	| 164.36006533558697 	|
+| xgboost 	            | 0.9986960112179987 	| 749.0485733114635 	|
+| lookback_xgboost 	    | 3.5651031645488147 	| 200.72583615414294 	|
+| bert_MA 	            | 10.83293233874024 	| 816.5899384610334 	|
 | news_rnn            	| 622.6200037110209  	| 741.4623210266183 	|
+| lookback_news_rnn 	| 15.457514085649745 	| 309.5794180019174 	|
+| sentiment_rnn 	    | 27.86876269801595 	| 1143.883111213618 	|
+| lookback_fasttext 	| 13.962031174946512 	| 169.93569939132175 	|
+
+</center>   
 
-</center>
+From the table, we can see that the `lookback_xgboost` model has the lowest RMSE value for the testing data, indicating that it is the best performing model among the four. The `bert_MA` model has the lowest RMSE value for the training data, but it has the highest RMSE value for the testing data, indicating that it may be overfitting the training data.
 
-From the table, we can see that the `univariate` model has the lowest RMSE value for the testing data, indicating that it is the best performing model among the four. The `bert_MA` model has the lowest RMSE value for the training data, but it has the highest RMSE value for the testing data, indicating that it may be overfitting the training data.
+Having news ruin the data dependency, ultimately they act as noise for the data. The `univariate` model, which only uses the stock prices, performs the best among the models. This suggests that the stock prices themselves contain enough information to predict future stock prices, and the addition of news headlines does not significantly improve the model's performance. 
 
-Having news ruin the data dependency, ultimately they act as noise for the data. The `univariate` model, which only uses the stock prices, performs the best among the models. This suggests that the stock prices themselves contain enough information to predict future stock prices, and the addition of news headlines does not significantly improve the model's performance.
 
+| Notebooks 	        | Loss Graph 	                        | Test Predictions	                |
+|:---:	                |:---:	                                |:---:	                            |
+| univariate 	        | ![alt text](images/image-2.png) 	    | ![alt text](images/image-6.png) 	|
+| xgboost 	            | 0.9986960112179987 	                | ![alt text](images/image-11.png)  |
+| lookback_xgboost 	    | 3.5651031645488147 	                | ![alt text](images/image-10.png)  |
+| bert_MA 	            | ![alt text](images/image.png) 	    | ![alt text](images/image-5.png) 	|
+| news_rnn            	| ![alt text](images/image-3.png)       | ![alt text](images/image-8.png) 	|
+| lookback_news_rnn 	| ![alt text](images/image-1.png) 	    | ![alt text](images/image-7.png) 	|
+| sentiment_rnn 	    | ![alt text](images/image-4.png)       | ![alt text](images/image-9.png) 	|
+| lookback_fasttext 	| ![alt text](images/image-12.png) 	    | ![alt text](images/image-13.png) 	|
 
diff --git a/images/image-1.png b/images/image-1.png
diff --git a/images/image-10.png b/images/image-10.png
diff --git a/images/image-11.png b/images/image-11.png
diff --git a/images/image-12.png b/images/image-12.png
diff --git a/images/image-13.png b/images/image-13.png
diff --git a/images/image-2.png b/images/image-2.png
diff --git a/images/image-3.png b/images/image-3.png
diff --git a/images/image-4.png b/images/image-4.png
diff --git a/images/image-5.png b/images/image-5.png
diff --git a/images/image-6.png b/images/image-6.png
diff --git a/images/image-7.png b/images/image-7.png
diff --git a/images/image-8.png b/images/image-8.png
diff --git a/images/image-9.png b/images/image-9.png
diff --git a/images/image.png b/images/image.png