This is just a timeline of various selected topics, source material and projects I was working on in the past to give you an overview and for me to keep track:
2020
Paper: TabNet
- TabNet: Attentive Interpretable Tabular Learning
- Keywords: Tabular data, interpretable neural networks, attention models
Project: Automated Stock Trading via Interactive Brokers
- Data Streaming via Interactive Brokers and RabbitMQ
- Data Storage via time series database TimeScaleDb
- Containerisation of Interactive Brokers Software
- Container orchestration for 5 workers:
- retrieving live stock data from Interactive Brokers API
- cleaning and storing stock data
- calculating time series features and stock indicators
- execution of trading strategy
- live trading of stocks via Interactive Brokers API
Multi-armed bandits
Paper: Causal Impact
- Inferring Causal Impact Using Bayesian Structural Time-Series Models
- Keywords: Causal inference, counterfactual, synthetic control, observational, difference in differences, econometrics, advertising, market research.
WhitePaper: Multiworld Testing:
- Multiworld Testing Decision Service: A System for Experimentation, Learning, And Decision-Making
- link
2019
Book: Forecasting: Principles and Practice
- Forecasting: Principles and Practice
Project: Prediction of Food Intolerances
Paper: Shap Values
- Consistent Individualized Feature Attribution for Tree Ensembles, Scott M. Lundberg et al.
- link
Paper: NgBoost
Framework: H2O
Recurrent Neural Networks: LSTMs
Bayesian AB testing
2018
Paper: Xgboost
Paper: Catboost
Paper: LightGbm
Project: Stock Trading Strategies
2017
Framework: D3 for Data Visualisation
Project: KnowNet Document Search Engine
- Development of a personal knowledge graph and graph based search engine that connects personal information (e.g. local or cloud files, emails and notes) in a network of nodes and allows contextual full text search
- WebApp development using React, Node.js and hosting on Heroku
- Business Model development
- Application for an Exist Business Start-up Grant
Python Packages for Natural Language Processing
- NLTK
- GENSIM
Paper: Topic Modelling
- Latent Dirichlet Allocation
- https://ai.stanford.edu/~ang/papers/nips01-lda.pdf
Paper: PageRank
2016
Paper: Monte Carlo Tree Search (MCTS)
- link
Frameworks:
- Apache Spark (Scala)
- Kafka (Scala)
Project: Big Data Risk Reporting Platform
- Design and Implementation for a DAX company
- Programming Language: Scala
Master Thesis:
- Modelling the global trade network: Nodes act as regional economical sectors (agriculture, automobile industry, etc.) and edges are trading routes between them
- This networks demand and supply can be perturbed by climatic extreme events (floods, droughts, tornados, etc.)
- Simulation of the distrubance propagation through the network
- Application of Decision Trees to predict the most vulnerable network nodes