Apple-Stock-Forecast-Agent

KO | EN

gitlite — search

#hacktoberfest #typescript #llm #android #deep-learning #linux #java #windows #awesome #kubernetes #vue #pytorch #ios #openai

Apple-Stock-Forecast-Agent

(★ 26)

No description available.

Apple-Stock-Forecast-Agent 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# ◈ AI Investment Agent

An intelligent investment assistant that combines multiple deep learning models (MLP, CNN-LSTM, Linear Regression) with Google Gemini AI to provide comprehensive stock analysis and actionable investment recommendations.

## Architecture

```
                         ┌─────────────────────────────┐
                         │       agent_system.py        │
  ┌──────────────┐       │                             │       ┌──────────────┐
  │              │       │  Tool 1: Price Predictor    │       │              │
  │   app.py     │─────▶│  Tool 2: Model Comparison   │─────▶│   Gemini     │
  │  (Streamlit) │       │  Tool 3: Technical Analysis │       │  2.5 Flash   │
  │              │◀─────│  Tool 4: Risk Assessment    │◀─────│              │
  └──────────────┘       └──────────────┬──────────────┘       └──────────────┘
                                        │
                         ┌──────────────▼──────────────┐
                         │    prediction_service.py    │
                         │                             │
                         │  ┌───────┐ ┌────────┐ ┌──┐  │
                         │  │  MLP  │ │CNN-LSTM│ │LR│  │
                         │  └───────┘ └────────┘ └──┘  │
                         │  + Technical Indicators     │
                         │  + Risk Engine              │
                         └─────────────────────────────┘
```

**Data flow:** User selects a model and asks a question → Agent runs all four tools (single-model prediction, cross-model comparison, technical analysis, risk assessment) → Gemini synthesizes a BUY / HOLD / SELL recommendation grounded in multi-model consensus.

## Models

| Model | Architecture | Input | Scaler | Strengths |
|---|---|---|---|---|
| **MLP** | Dense feed-forward | Close price window (flat) | MinMaxScaler | Fast, stable on blue-chip patterns |
| **CNN-LSTM** | Conv1D → LSTM stack | 60-step × 38 features (3D) | StandardScaler | Highest capacity, captures complex temporal + spatial patterns |
| **Linear Regression** | OLS | 30-day close window | MinMaxScaler | Interpretable baseline for comparison |

The frontend overlays all three forecasts on a single chart and displays a consensus badge (ALL BULLISH / ALL BEARISH / MIXED) so users can gauge model agreement at a glance.

## Core Components

| File | Role |
|---|---|
| `prediction_service.py` | Loads all models, runs predictions, computes indicators and risk metrics. Exposes `active_model` switching and `compare_models()`. |
| `agent_system.py` | Defines four analysis tools, runs them, and sends combined output to Gemini for structured reasoning. |
| `app.py` | Streamlit dashboard — model selector, metric cards, comparison table, multi-line forecast chart, and natural-language Q&A. |

## Quick Start

### 1. Prerequisites
- Python 3.9+
- `mlp_model_final.h5` (trained MLP weights)
- `stock_prediction_cnn_lstm_plots/apple_stock_cnn_lstm_model.h5` (trained CNN-LSTM weights)
- `apple_5yr_daily.csv` (historical AAPL data)
- Google API key for Gemini

### 2. Install
```bash
pip install -r requirements.txt
```

### 3. Configure
Create `.env` in the project root:
```env
GOOGLE_API_KEY=your_key_here
```

### 4. Run
```bash
streamlit run app.py
```
Open `http://localhost:8501`.

## 📁 Project Structure
ai-investment-agent/
│
├── prediction_service.py      # Multi-model prediction service
├── agent_system.py            # AI Agent with Google Gemini
├── app.py                     # Streamlit web interface
│
├── Data/                      # Stock data
│   └── apple_5yr_daily.csv       # 5 years of AAPL data
│
├── mlp_model_final.h5         # MLP neural network model
├── cnn_lstm_model.h5          # CNN-LSTM hybrid model (optional)
├── linear_model_final.h5      # Linear regression model (optional)
│
├── .env                       # API keys (create this, not in Git)
├── requirements.txt           # Python dependencies
└── README.md                  # This file

## Analysis Pipeline
For every user query the agent runs four tools:
1. **Price Predictor** — forecasts the next N days using the currently selected model (MLP, CNN-LSTM, or Linear).
2. **Model Comparison** — runs all available models and reports trend, change %, and confidence for each, plus a consensus signal.
3. **Technical Analysis** — RSI, MACD crossover, moving averages, and Bollinger Band positioning.
4. **Risk Assessment** — annualised volatility, max drawdown, composite risk score, and position-sizing advice.
Gemini receives all four outputs and returns a structured recommendation with reasoning, position size, and stop-loss rules.

## 🛠️ Technology Stack

| Category | Technologies |
|----------|-------------|
| **Deep Learning** | TensorFlow/Keras (MLP, CNN-LSTM) |
| **Machine Learning** | scikit-learn (Linear Regression) |
| **AI Agent** | Google Gemini 2.5 Flash |
| **Data Processing** | pandas, NumPy, MinMaxScaler |
| **Frontend** | Streamlit, Plotly |
| **Environment** | python-dotenv |

## Requirements

```
streamlit>=1.30.0
plotly>=5.18.0
pandas>=2.0.0
numpy>=1.24.0
scikit-learn>=1.3.0
tensorflow>=2.14.0
python-dotenv>=1.0.0
google-genai>=1.0.0
```

## Disclaimer

This project is for **educational and demonstration purposes only**. It does not constitute financial advice. Stock predictions are inherently uncertain — never make investment decisions based solely on model outputs.

Apple-Stock-Forecast-Agent

Apple-Stock-Forecast-Agent 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Apple-Stock-Forecast-Agent

Apple-Stock-Forecast-Agent 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README