DishDash - Machine Learning Team

Team Member:

Machine Learning members of DishDash consist of:

• Yeftha Joshua Ezekiel M006BSY0403
• Raffel Prama Andhika M006BSY0185
• Labiba Adinda Zahwana M200BSX0414

Importing Libraries:

import tensorflow as tf
import numpy as np
import pandas as pd
import keras
from sklearn.metrics.pairwise import cosine_similarity
from sklearn.preprocessing import LabelEncoder
import random

This code imports the TensorFlow, NumPy, Pandas, Keras, Random, and Skelearn libraries, which are commonly used for machine learning and numerical computations.

Installing Libraries:

!pip install [library-name]

These lines use the !pip install command to install Python libraries.

Saving the Trained Model:

model.save("model.h5")

This line saves the trained model as an HDF5 file named "model.h5". The model was likely trained using a deep learning framework like TensorFlow.

Dataset:

- dot_prod.csv                       # Dot product result for content-based model
- final_food.csv                     # Food dataset containing nutrition and image
- food_resto_location.csv            # Dataset containing food name, restaurant, and restaurant location
- like_data.csv                      # User dataset containing user id, food id, and like
- preprocess_content_based_data.csv  # Dataset for food label
- resto_location.csv                 # Raw dataset for restaurant location

Model Notebook:

• Content-Based_Model.ipynb --> Notebook for content based recommender system model
• Like_Based_Model.ipynb --> Notebook for like based recommender system model

Result:

The model attained a training accuracy of 97.78% on the training dataset and 97.75% on the test dataset. Additionally, the model utilizes the AUC (Area Under the Receiver Operating Characteristic Curve), achieving a value of 0.937 on the training dataset and 0.933 on the test dataset.