Healthcare & Finance AI

import tensorflow as tf
from tensorflow import keras
import pydicom
import numpy as np

class MedicalImageClassifier:
    def __init__(self, model_path):
        self.model = keras.models.load_model(model_path)
        self.preprocessing = self.setup_preprocessing()
    
    def diagnose_chest_xray(self, dicom_path):
        # Load DICOM file
        ds = pydicom.dcmread(dicom_path)
        image = ds.pixel_array
        
        # Preprocess image
        image = self.preprocess_medical_image(image)
        
        # Make prediction
        predictions = self.model.predict(np.expand_dims(image, axis=0))
        
        # Interpret results
        conditions = ['Normal', 'Pneumonia', 'COVID-19', 'Tuberculosis']
        probabilities = tf.nn.softmax(predictions[0])
        
        results = {
            condition: float(prob) 
            for condition, prob in zip(conditions, probabilities)
        }
        
        # Add confidence score and explanation
        max_prob = max(results.values())
        diagnosis = max(results, key=results.get)
        
        return {
            'diagnosis': diagnosis,
            'confidence': max_prob,
            'all_probabilities': results,
            'requires_review': max_prob < 0.85
        }
    
    def preprocess_medical_image(self, image):
        # Normalize pixel values
        image = image.astype(np.float32) / 255.0
        # Resize to model input size
        image = tf.image.resize(image, (224, 224))
        # Apply CLAHE for contrast enhancement
        return self.apply_clahe(image)

import pandas as pd
import numpy as np
from sklearn.ensemble import RandomForestClassifier
import yfinance as yf

class AITradingStrategy:
    def __init__(self, symbol, lookback=20):
        self.symbol = symbol
        self.lookback = lookback
        self.model = RandomForestClassifier(n_estimators=100)
        
    def prepare_features(self, data):
        # Technical indicators
        data['SMA_20'] = data['Close'].rolling(window=20).mean()
        data['SMA_50'] = data['Close'].rolling(window=50).mean()
        data['RSI'] = self.calculate_rsi(data['Close'])
        data['MACD'] = self.calculate_macd(data['Close'])
        
        # Price features
        data['Returns'] = data['Close'].pct_change()
        data['Volatility'] = data['Returns'].rolling(window=20).std()
        
        # Volume features
        data['Volume_Ratio'] = data['Volume'] / data['Volume'].rolling(window=20).mean()
        
        return data
    
    def generate_signals(self, current_data):
        features = self.prepare_features(current_data)
        
        # Make prediction
        X = features[['SMA_20', 'SMA_50', 'RSI', 'MACD', 'Volatility', 'Volume_Ratio']].iloc[-1:]
        prediction = self.model.predict_proba(X)[0]
        
        # Generate trading signal
        signal = {
            'action': 'BUY' if prediction[1] > 0.6 else 'SELL' if prediction[0] > 0.6 else 'HOLD',
            'confidence': max(prediction),
            'risk_score': self.calculate_risk(features),
            'position_size': self.calculate_position_size(prediction, features)
        }
        
        return signal
    
    def calculate_risk(self, features):
        # Value at Risk calculation
        returns = features['Returns'].dropna()
        var_95 = np.percentile(returns, 5)
        return abs(var_95)