Implementation of Multilayer Feed-Forward Neural Network (MLP) on Wine Quality Dataset using Keras

 

Experiment

Implementation of Multilayer Feed-Forward Neural Network (MLP) on Wine Quality Dataset using Keras

---

🎯 Objective

• To implement a MLP using Keras
• To train and evaluate the model on the Wine Quality dataset
• To understand:
  • Forward propagation
  • Backpropagation  
  • Model evaluation

---

Theory

---

🔹 1. Multilayer Feed-Forward Network (MLP)

• A type of artificial neural network
• Consists of:
  • Input layer
  • One or more hidden layers
  • Output layer

👉 Data flows only in one direction (no cycles)

---

🔹 2. Wine Quality Dataset

• Contains chemical properties of wine:
  • acidity, sugar, pH, alcohol, etc.
• Output:
  • Wine quality score (0–10)

👉 Can be treated as:

• Classification problem (recommended for lab)

---

🔹 3. Forward Propagation

$$Z = XW + b$$
$$A = f(Z)$$

---

🔹 4. Backpropagation

• Computes gradients using chain rule
• Updates weights to minimize loss

---

🔹 5. Activation Functions

• Hidden layers → ReLU
• Output layer → Softmax

---

🔹 6. Loss Function

• Categorical Crossentropy

---

🔹 7. Optimizer

• Adam (adaptive gradient method)

---

💻 Program (Keras Implementation)

---

📌 Step 1: Import Libraries

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import LabelEncoder

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Input
from tensorflow.keras.utils import to_categorical

---

📌 Step 2: Load Dataset

# Load dataset (update path if needed)
data = pd.read_csv("/content/WineQT.csv", sep=',')

print(data.head())

---

📌 Step 3: Preprocessing

# Features and target
X = data.drop('quality', axis=1).values
y = data['quality'].values

# Convert labels to categorical
encoder = LabelEncoder()
y = encoder.fit_transform(y)
y = to_categorical(y)

# Train-test split
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=0)

# Feature scaling
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

---

📌 Step 4: Build MLP Model

model = Sequential([
    Input(shape=(X_train.shape[1],)),
    Dense(32, activation='relu'),
    Dense(16, activation='relu'),
    Dense(y.shape[1], activation='softmax')
])

---

📌 Step 5: Compile Model

model.compile(optimizer='adam',
              loss='categorical_crossentropy',
              metrics=['accuracy'])

---

📌 Step 6: Train Model

history = model.fit(X_train, y_train,
                    epochs=150,
                    batch_size=100,
                    validation_split=0.2,
                    verbose=1)

---

📌 Step 7: Evaluate Model

loss, accuracy = model.evaluate(X_test, y_test)
print("Test Accuracy:", accuracy)

---

📈 Loss and Accuracy Visualization

# Loss
plt.plot(history.history['loss'], label='train loss')
plt.plot(history.history['val_loss'], label='val loss')
plt.legend()
plt.title("Loss Curve")
plt.show()

# Accuracy
plt.plot(history.history['accuracy'], label='train acc')
plt.plot(history.history['val_accuracy'], label='val acc')
plt.legend()
plt.title("Accuracy Curve")
plt.show()

---

🔍 Observations

• Training loss decreases over epochs
• Validation accuracy stabilizes
• Model learns meaningful patterns

---

Result

• MLP successfully classifies wine quality
• Achieves reasonable accuracy depending on training