Overview

The Machine Learning (ML) module for the HighKing project is designed to provide personalized hiking trip and mountain recommendations. This module leverages advanced algorithms and extensive datasets to ensure that users receive the best possible suggestions based on their preferences and behaviors.

Features

• Personalized Trip Recommendations: Suggests hiking trips tailored to user preferences.
• Mountain Recommendations: Suggests mountain recommendation tailored to user preferences.

ML Models

Trip Recommendation Model

• Architecture:
  • Dense(128, activation='relu', input_shape=(features.shape[1],), kernel_regularizer=l2(0.01))
  • BatchNormalization()
  • Dropout(0.5)
  • Dense(64, activation='relu', kernel_regularizer=l2(0.01))
  • BatchNormalization()
  • Dropout(0.5)
  • Dense(32, activation='relu', kernel_regularizer=l2(0.01))
  • BatchNormalization()
  • Dense(features.shape[1], activation='selu')
• Performance:
  • Loss: 0.0695
  • MAE: 0.1414
  • Validation Loss: 0.0851
  • Validation MAE: 0.1355

Mountain Recommendation Model

• Architecture:
  • Dense(128, activation='relu', input_shape=(X_features.shape[1],))
  • Dropout(0.5)
  • Dense(64, activation='relu')
  • Dropout(0.5)
  • Dense(X_features.shape[1], activation='sigmoid')
• Performance:
  • Loss: 0.0435
  • MAE: 0.0536
  • Validation Loss: 0.0793
  • Validation MAE: 0.0689

Common Techniques

• Cosine Similarity: Used to measure similarity between user preferences and available trip options.
• TensorFlow: Utilized for building and training the models.

Data Processing

Data Collection

• Mountain Dataset: Scraped from Google for comprehensive mountain information.
• Trip Information: Crawled from various sources to gather extensive trip data.