Giter Club home page Giter Club logo

deepbake's Introduction

DeepBake

Baking Machine Learning into Great British Bake Off

What is this?

DeepBake is a set of deep learning neural network models to predict the final rankings of GBBO contestants.

DeepBake consists of 10 models for each episode, and was trained on data from seasons 2-9.

Data include 8 variables:

  • Technical Challenge Ranking for that week and running mean from prior weeks
  • Contestant was Star Baker and running mean of times named Star Baker
  • Contestant was a favorite baker that week and running mean from prior weeks
  • Contestant was an unfavored baker that week and the running mean

Data were obtained from Wikipedia. Thanks to those who made those pages.

Data were then quantile scaled to fit a normal distribution.


Does this work?

DeepBake's performance was measured using a Leave One Out method. One season was set aside for evaluation while training the model on using the remaining seasons. A mean receiver operating curve was calculated by iterating through all seasons.

The closer the area under the curve (AUC) is to 1, the more accurate the model.

Random chance of making a correct prediction has an AUC of 0.5 (dotted diagonal line). The Episode 4 model has an AUC of 0.91 (+/- 0.04 95% Confidence Interval), meaning it has a very good chance of predicting the final GBBO winner!

DeepBake makes 5 predictions:

  • 1st Place ๐Ÿ† ๐Ÿ†
  • Runner-Up ๐Ÿ†
  • 3rd-4th Place
  • 5th-7th Place
  • 8th Place and Below

The evaluation was measured using this tiered class system.

Note how the classifier gets better at predicting as the season progresses. This makes sense because the good bakers rise to the top (favored and star bakers) and historical data are recorded as running means.


Does this mean DeepBake can predict the winner for Season 10?

Absolutely! Here are the current standings!

Season 10 : Episode 2 Predictions

Finalist Prediction

DeepBake puts Alice in the lead with a 36.8% probability score for being the finalist. Michael, David, and Rosie are close contenders with around 21% probability.

Finalist + Runner-Up

This score is the addition of the finalist probability and the runner-up probability. It's a measurement of how likely a baker would be in the final episode.

DeepBake thinks Alice (87%), David (62%), and Michael (60%) will vie for the title of best baker.

8th and Below

DeepBake gave Dan and Jamie the highest scores (80% and 78%) for being in the bottom tier. Dan was eliminated in week 1, while Jamie was eliminated at the end of episode 2.

In fact, DeepBake can make a prediction before the judges eliminate a baker. These results suggest DeepBake correctly predicted Jamie would leave the tent!


Stay tuned for Week 3!

deepbake's People

Contributors

dantaki avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.