Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,77 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
---
|
6 |
+
|
7 |
+
🔍 Model Overview
|
8 |
+
Feature Description
|
9 |
+
📦 Model Type LSTM (Long Short-Term Memory), a type of Recurrent Neural Network (RNN)
|
10 |
+
🧠 Frameworks Used TensorFlow (Keras API), Scikit-learn, yfinance, NumPy, Pandas
|
11 |
+
📈 Input Past 60 days of Bitcoin closing prices
|
12 |
+
🎯 Output The predicted closing price for the next day
|
13 |
+
🧪 Goal Short-term (1-day ahead) price prediction
|
14 |
+
📊 Evaluation Metric RMSE (Root Mean Squared Error)
|
15 |
+
|
16 |
+
🔧 What It Does
|
17 |
+
Downloads historical BTC-USD data from Yahoo Finance
|
18 |
+
|
19 |
+
Normalizes the closing price data between 0 and 1 using MinMaxScaler
|
20 |
+
|
21 |
+
Splits the data into 80% training and 20% testing
|
22 |
+
|
23 |
+
Uses a sliding window of 60 days to create time-sequenced input for LSTM
|
24 |
+
|
25 |
+
Builds a 2-layer LSTM network with dropout to avoid overfitting
|
26 |
+
|
27 |
+
Trains the model to minimize mean squared error
|
28 |
+
|
29 |
+
Evaluates the model on test data
|
30 |
+
|
31 |
+
Plots actual vs predicted prices
|
32 |
+
|
33 |
+
Makes a future prediction for the next day based on the last 60 days
|
34 |
+
|
35 |
+
💡 What This Model Is Useful For
|
36 |
+
Educational purposes (learning LSTM, time series preprocessing)
|
37 |
+
|
38 |
+
Research and experimentation with financial prediction models
|
39 |
+
|
40 |
+
Visualization and evaluation of model performance
|
41 |
+
|
42 |
+
Academic paper support (as you’re already writing)
|
43 |
+
|
44 |
+
⚠️ Limitations
|
45 |
+
It uses only closing price – no volume, sentiment, or indicators like RSI/MACD
|
46 |
+
|
47 |
+
It performs single-step prediction (only the next day)
|
48 |
+
|
49 |
+
May not adapt well to sudden market shocks
|
50 |
+
|
51 |
+
Not suitable as-is for high-frequency trading or live trading bots
|
52 |
+
|
53 |
+
🚀 How to Improve It
|
54 |
+
Include more features: trading volume, high/low prices, technical indicators
|
55 |
+
|
56 |
+
Add external data: news sentiment, macroeconomic indicators
|
57 |
+
|
58 |
+
Use attention mechanisms or transformer architectures
|
59 |
+
|
60 |
+
Convert it to multi-step forecasting (predict next 3, 5, or 7 days)
|
61 |
+
|
62 |
+
Wrap it in a REST API or real-time dashboard
|
63 |
+
|
64 |
+
Let me know if you want to:
|
65 |
+
|
66 |
+
Package this into a web app
|
67 |
+
|
68 |
+
Prepare a GitHub README for the code
|
69 |
+
|
70 |
+
Extend the model for live trading (e.g., using Binance API)
|
71 |
+
|
72 |
+
Or build a production-ready pipeline
|
73 |
+
---
|
74 |
+
⚠️ ⚠️ ⚠️ ⚠️ ⚠️ ⚠️ ⚠️ ⚠️ ⚠️ ⚠️
|
75 |
+
Disclaimer:
|
76 |
+
This model is developed strictly for experimental and educational purposes. It is not intended for real-world financial forecasting or investment decisions. Users are strongly advised not to rely on the outputs of this model for any form of trading or financial transactions. No guarantees are made regarding the accuracy or reliability of the predictions, and no liability or responsibility is accepted for any losses or damages resulting from the use of this system. By using this model, you acknowledge and agree that no claims or disputes will be accepted under any circumstances.
|
77 |
+
---
|