Kaggle Challenge: TalkingData AdTracking Fraud Detection

Discover Hao Chen's feedback on his first Kaggle Challenge

By Hao Chen Data Scientist


This article was written after finishing my first Kaggle challenge – TalkingData AdTracking Fraud Detection Challenge to share what I’ve learnt from this challenge. In this article, I’ll begin with a brief introduction to Kaggle and to TalkingData AdTracking Fraud Detection Challenge, then I will talk about what I’ve learnt after this challenge.


Kaggle Platform



According to Wikipedia,

Kaggle is a platform for predictive modelling and analytics competitions in which statisticians and data miners compete to produce the best models for predicting and describing the datasets uploaded by companies and users. This crowdsourcing approach relies on the fact that there are countless strategies that can be applied to any predictive modelling task and it is impossible to know beforehand which technique or analyst will be most effective.

Recently, people added more functionalities to Kaggle, such as sharing your code in “Kernels”, asking a question in “Discussions”, learning a new Data Science technique in “Learn” and finding your job in “Jobs” etc. Kaggle has became the # 1 platform for data scientists and machine learners.


TalkingData AdTracking Fraud Detection Challenge


Fraud risk is everywhere, but for companies that advertise online, click fraud can happen at an overwhelming volume, resulting in misleading click data and wasted money. Ad channels can drive up costs by simply clicking on the ad at a large scale. With over 1 billion smart mobile devices in active use every month, China is the largest mobile market in the world and therefore suffers from huge volumes of fradulent traffic.

TalkingData, China’s largest independent big data service platform, covers over 70% of active mobile devices nationwide. They handle 3 billion clicks per day, of which 90% are potentially fraudulent. Their current approach to prevent click fraud for app developers is to measure the journey of a user’s click across their portfolio, and flag IP addresses who produce lots of clicks, but never end up installing apps. With this information, they’ve built an IP blacklist and device blacklist.

While successful, they want to always be one step ahead of fraudsters and have turned to the Kaggle community for help in further developing their solution. In their 2nd competition with Kaggle, you’re challenged to build an algorithm that predicts whether a user will download an app after clicking a mobile app ad. To support your modeling, they have provided a generous dataset covering approximately 200 million clicks over 4 days!


Steps of finishing a challenge


There are 4 crucial steps with their importances:

  • 80% feature engineering
  • 10% making local validation as fast as possible
  • 5% hyper parameter tuning
  • 5% ensembling

Feature engineering


The most important step for winning a challenge is the feature engineering which can take up 80% importances. For a better prediction, you need add up all features that you can find or compose from your original dataset, such as Group by, unique, count, cumulative_count, sort, shift(next & previous), mean, variance, etc.


Making local validation


The next step is to make a local validation set as fast as possible. Once it is created, you can tune your hyperparameters and test your feature importances on your local validation set.

If you don’t have a good validation scheme then you rely solely on LB probing, which can easily lead to overfit.


Hyper parameter tuning


Two most used methods are XGBoost and Light GBM, you can refer to this website xgboost/Light GBM parameters to set up all your parameters quickly.

To get a better accuracy, you need to tune your parameters with Gird Search. For example:

param_grid = {'n_estimators': [300, 500], 'max_features': [10, 12, 14]}

model = grid_search.GridSearchCV(estimator=rfr, param_grid=param_grid, n_jobs=1, cv=10, verbose=20, scoring=RMSE), y_train)



It is more and more difficult to get good predictions with just a single model. If we can combine several medels, the prediction result could usually be better.

There are some ensemble methods:

  • Bagging: Each Base Model is trained using different random subsets of the training data, and the final results are voted by each base model with the same weight. That is the principle of Random Forest.
  • Boosting: The Base Model is trained iteratively, each time modifying the weight of the training sample based on the prediction errors in the previous iteration. This is also the principle of Gradient Boosting. Better than Bagging but easier to Overfit.
  • Blending: Different Base Models are trained with disjoint data and their outputs are averaged (weighted). Simple implementation, but less use of training data.
  • Stacking: The whole process is very much like Cross Validation. First, the training data is divided into 5 parts, followed by a total of 5 iterations. At each iteration, 4 data are trained as Training Sets for each Base Model, and then the remaining Hold-out Set is used for prediction. At the same time, its predictions on test data should also be preserved. In this way, each Base Model will make a prediction of one of the training data at each iteration and make a prediction of all the test data. After 5 iterations have been completed, we have obtained a matrix of # training data rows x #Base Model number. This matrix is ​​then used as the training data of the second-level Model. After the second-level Model is trained, the previously saved Base Model predicts the test data (because each Base Model has been trained five times, and the entire test data has been predicted five times, so for these five requests An average value, which results in a matrix with the same shape as the second-level training data) is taken out for prediction and the final output is obtained.



Kaggle is possibly the best platform for data scientists to practice their skills. To start with this challenge, I got inspired from kernels and discussions. Then I tried Light GBM and XGBoost that I had never used before. I’ll continue fighting on Kaggle!




  1. TOP5%Kaggler: How to get into top 10% ranking(In Chinese)
  2. Solution #6 overview
  3. xgboost/Light GBM parameters



Hao Chen ranked in the top 8% for this challenge (284/3967 players).

By Hao Chen Data Scientist
Data to business

Successfully completing a data project: a path still strewn with pitfalls

In 2020, corporate investment in data projects is expected to exceed 203 billion dollars worldwide. But at a time when many are claiming to be Data Driven Companies, lots of data projects end in failure. Yet most of these failures are unnecessary and due to well-known causes! Focus on the recurrent pitfalls to avoid.

05/02/2019 Discover 

Data Science applied to the retail industry: 10 essential use cases

Data Science is having an increasing impact on business models in all industries, including retail. According to IBM, 62% of retailers say the use of Big Data techniques gives them a serious competitive advantage. Knowing what your customer wants and when, is today at your fingertips thanks to data science. You just need the right tools and the right processes. We present in this article 10 essential applications of data science in the field of retail.

31/05/2018 Discover 

Introduction to TensorFlow on the datalab of Google Cloud Platform

TensorFlow is a software library, open source since 2015, of numerical computation developed by Google. The particularity of TensorFlow is its use of data flow graphs.

30/05/2018 Discover 

Lamport clocks and the pattern of the Idempotent Producer (Kafka)

Do you know the Lamport clocks? Devoxx France 2018 was the opportunity, during the very interesting talk of DuyHai DOAN , to discover or rediscover this algorithm formalized by Leslie Lamport in 1978, more than ever used today in the field of distributed systems, and which would have inspired the Kafka developers in the implementation of the pattern of Idempotent Producer .

23/05/2018 Discover 

Establishment of a centralised log management platform with the Elastic suite

The volume of data generated by our systems and applications continues to grow, resulting in the proliferation of data centers and data storage systems.  In the face of this data explosion and the investment in skills and resources, decision-makers need sophisticated analysis and sophisticated dashboards to help them manage their systems and customers.

14/05/2018 Discover 

Introduction to Reactive Systems

Margo Consultants participated in  Devoxx France 2018 , the conference for Passionate Developers, organized from April 18 to 20, 2018 in Paris. Discover a synthesis on reactive systems illustrated by a concrete use case.

11/05/2018 Discover