Using Machine Learning to Guess What's Cooking

This is for UIC CS412 course project.

Team Name:

AllEatable

Team members (Foodies):

• Huang, He
• Liu, Ye
• Sun, Lichao
• Wang, Zhu
• Xia, Congying
• Zhu, Fan

We are using this Kaggle Competition as our project topic.

Important Deadlines:

Mar 16, proposal due    
Apr 4, progress report due  
Apr 25: presentations
Apr 27, presentations, final report due 

Problem definition:

1.Text Classification：
    1) Features: all V ingredients, and each recipe is represented as a vector of length V.
    2) Classes: all K kinds of cuisines 

ML Algorithms:

1. SVM (multiple classes version)
2. Random Forest
3. Naive Bayes
4. Class-baed Association Rule Mining
5. Neural Network
6. Latent Dirichlet Allocation

Visualization:

1.reduce V-dimensional vectors to 2- or 3- dimensional vectors
2. Methods:
    1) PCA
    2) t-SNE        

t-SNE example

Contributions:

1. Using many traditional ML techniques
2. Compare the performance of each used technique
3. Visualizing the recipes in 2-D/3-D space to show the closeness of recipes.