About

Connect with me

Linkedin: https://www.linkedin.com/in/chris-dinant/
Kaggle: https://www.kaggle.com/chrisdinant
Github: https://www.kaggle.com/chrisdinant

Comments

Post a Comment

Popular posts from this blog

Kaggle Tensorflow Speech Recognition Challenge

Image
Implementation of the ResNet and CTC models at https://github.com/chrisdinant/speech In November of 2017 the Google Brain team hosted a speech recognition challenge on Kaggle . The goal of this challenge was to write a program that can correctly identify one of 10 words being spoken in a one-second long audio file. Having just made up my mind to start seriously studying data science with the goal of turning a new corner in my career, I decided to tackle this as my first serious kaggle challenge. In this post I will talk about ResNets, RNNs, 1D and 2D convolution, Connectionist Temporal Classification and more. Let's go! Exploratory Data Analysis The training data supplied by Google Brain consists of ca. 60,000 1-second-long .wav files in 32 directories that are named by the word spoken in the files. Only 10 of these are classes you need to identify, the others should go in the 'unknown' or 'silence' classes. There are a couple of things you can do to get a...
Read more

Post-Script to the Purpose of this Blog

Image
I said in my first post that I have this blog in the hope that it is useful to other people just starting to learn DataScience. This is not really to whole truth. The other reason, and perhaps more importantly, is that you learn so much from trying to explain things to other people. Who was it that said that you don't really understand something unless you can explain it to a five-year-old? Was it Feynman? That's really what I'm doing here. Making sure that I really understand the concepts and techniques that I'm learning and not just storming through course after course and book after book without the real end-goal in mind: become a better data scientist.
Read more

Bayes and Binomial Theorem

Image
Bayes Theorem In statistics there are many situations where you want to determine the probability that a sample for which you have certain measurement belongs to a certain set. Say you want to know the chance that you have HIV if you test positive. No test is perfect, so this probability will depend on the test sensitivity, but also on the specificity and on the incidence in the population, or set, that you belong to. Bayes Theorem is a simply the logic you have to apply to estimate such probabilities. As a cancer researcher my attention was naturally drawn to this paper currently trending on Pubmed: Detection and localization of surgically resectable cancers with a multi-analyte blood test. This is a perfect practical example for applying Bayes rule! And most of the information we need is right there in the abstract: " The sensitivities ranged from 69% to 98% for the detection of five cancer types (ovary, liver, stomach, pancreas, and esophagus)... " and " The ...
Read more