Displaying 20 to 40 from 47 results

kaggle-lshtc - Code for Large Scale Hierarchical Text Classification competition. Final place: 3rd

  •    C++

Code for Large Scale Hierarchical Text Classification competition. a centroid-based flat classifier.

kaggle-ndsb - Code for National Data Science Bowl. 10th place.

  •    Lua

Code for National Data Science Bowl at Kaggle. Ranked 10th/1049. Ensemble Deep CNNs trained with real-time data augmentation.

kaggle-ndsb2 - Code for Second Annual Data Science Bowl. 16th place.

  •    Lua

Code for Second Annual Data Science Bowl. 16th place. A Hybrid Deep Neural Network using CNN and MLP.

Apartment-Interest-Prediction - Predict people interest in renting specific NYC apartments

  •    Jupyter

Predict people interest in renting specific apartments. The challenge combines structured data, geolocalization, time data, free text and images. This solution features Gradient Boosted Trees (XGBoost and LightGBM) and does not use stacking, due to lack of time.




Kaggle-MNIST - Simple ConvNet to classify digits from the famous MNIST dataset

  •    Python

Simple ConvNet to classify digits from the famous MNIST dataset. This program gets 98.63% on Kaggle's test set. In order to run this program, you need to have Theano, Keras, and Numpy installed as well as the train and test datasets (from Kaggle) in the same folder as the python file.

March-Madness-2017 - Kaggle Competition for Predicting NCAA Basketball Tourney Games

  •    Jupyter

Kaggle Competition for Predicting NCAA Basketball Tourney Games. Link to the associated blog post I wrote.

kaggle-carvana - Solution for the Carvana Image Masking Challenge on Kaggle

  •    Python

The solution for the Carvana Image Masking Challenge on Kaggle. It uses a custom version of RefineNet with Squeeze-and-Excitation modules implemented in PyTorch. It was a part of the final ensemble that was ranked 23 out of 735 teams (top 4%). The goal of the Carvana Image Masking Challenge was to develop an algorithm that removes a background from a wide variety of car photos. Here you can see predictions from a trained neural network for 16 images of a single car.


ds_bowl_2018 - Kaggle Data Science Bowl 2018

  •    Jupyter

This is a DWT-inspired solution to the Kaggle's 2018 DS Bowl I produced within approximately 1 week before the end of the compeititon. UPDATE 2018-04-22 - my score was 114th. I guess they are cleaning the LB in the end.

open-solution-avito-demand-prediction - Open solution to the Avito Demand Prediction Challenge

  •    Jupyter

This is an open solution to the Avito Demand Prediction Challenge. In this open source solution you will find references to the neptune.ml. It is free platform for community Users, which we use daily to keep track of our experiments. Please note that using neptune.ml is not necessary to proceed with this solution. You may run it as plain Python script 😉.

open-solution-cdiscount-starter - Open solution to the Cdiscount’s Image Classification Challenge

  •    Python

This is ready to use, end-to-end sample solution for the currently running Kaggle Cdiscount challenge. It involves data loading and augmentation, model training (many different architectures), ensembling and submit generator.

open-solution-data-science-bowl-2018 - Open solution to the Data Science Bowl 2018

  •    Python

This is an open solution to the Data Science Bowl 2018 based on the topcoders winning solution from ods.ai. In this open source solution you will find references to the neptune.ml. It is free platform for community Users, which we use daily to keep track of our experiments. Please note that using neptune.ml is not necessary to proceed with this solution. You may run it as plain Python script 😉.

open-solution-ship-detection - Open solution to the Airbus Ship Detection Challenge

  •    Python

This is an open solution to the Airbus Ship Detection Challenge. In this open source solution you will find references to the neptune.ml. It is free platform for community Users, which we use daily to keep track of our experiments. Please note that using neptune.ml is not necessary to proceed with this solution. You may run it as plain Python script 🐍.

open-solution-talking-data - Open solution to the TalkingData AdTracking Fraud Detection Challenge

  •    Jupyter

This is an open solution to the TalkingData Challenge. Deliver open source, ready-to-use and extendable solution to this competition. This solution should - by itself - establish solid benchmark, as well as provide good base for your custom ideas and experiments.

open-solution-toxic-comments - Open solution to the Toxic Comment Classification Challenge

  •    Python

Here, at Neptune we enjoy participating in the Kaggle competitions. Toxic Comment Classification Challenge is especially interesting because it touches important issue of online harassment. You need to be registered to neptune.ml to be able to use our predictions for your ensemble models.

home-credit-default-risk - Default risk prediction for Home Credit competition - Fast, scalable and maintainable SQL-based feature engineering pipeline

  •    Python

This is code I built for the Home Credit default risk competition on Kaggle. This should be seen more as an ML engineering achievement than a data science top of the line prediction model. First of all, due to time constraints this is not a top scorer. First rank was 0.80570 AUC (499 submissions), this is 0.78212 AUC (12 submissions).

Credit-Card-Fraud - A Very Deep Neural Network that can classify Credit Card Fraudulent Transaction with 99

  •    Jupyter

In the world of Technology, Credit Card Fraudulent Transactions are fairly Common and this is a Deep Neural Network (Algorithm) that can classify these transactions just by looking at the data with 99.92% Accuracy which is likely to be very accurate. This Neural Network is based of the Credit-Card-Fraud Data available on Kaggle, It contains a whooping 248,407 Transactions which occurred in September 2013 by European Card Holders.