Recommender System is the name given to any software designed to recommend you objects to purchase, to click on or to watch, which might be attractive to you with a high probability. Development of such systems became a hot topic for software developers and machine learning engineers in the last years. A well-performing recommender system can make significant contributions to various online platforms.

Nowadays, many of us interact with recommender systems on a daily basis. If you have an account in Amazon or Netflix or many similar online services, you may have received suggestions that often fit your interest. This can vary from a movie suggested by Netflix or very various products suggested by Amazon. The targeted advertisements on Facebook or the targeted pop-up advertisements in your browser are also other examples of applications of these systems. Almost no two person receives the exact same advertisements or suggestions on an online platform.  All these personalized suggestions and advertistments use recommender systems to find the best match for your interests according to your former ratings, purchases or clicks.

Content-based recommender systems often make use of the explicit features of the users and the items to provide recommendations. In contrary, the Collaborative Filtering recommender systems do not have access to any sort of explicit features and tend to learn the important features of users and items implicitly. The real-world recommender systems are often a hybrid of both approaches.

You can try our movie recommender system demo, by simply giving your own ratings and see what the system will suggest to you. This movie recommender system is based on the implementation of Matrix Factorization algorithm in Python. Matrix Factorization is a collaborative filtering approach which tends to learn implicit features for users and items by factorizing the exiting rating matrix. We expalin this algorithm in more details, in this Google Colaboratory Notebook . We describe how the recommender system problem turns into an optimization problem and how to handle such optimization tasks. Also a step-by-step implementation of this algorithm in Python is provided.

One of the slides of the presentation given by Henning Buhl during the Digital Xchange 2019 Conference

On Saturday, May, 25th, 2019, TH Köln hosted the annual Digital Xchange conference, which is organized as a cooperative work of Opitz Consulting and TH Köln, for the third time. This 1-day conference is becoming more and more popular: this year it attracted more than 800 participants.  More than 100 talks were given in the field of digitalization and its impact on industry. The topics had a large variety and covered many areas of digitalization including Big Data, Cloud Computing, Industry 4.0, Machine Learning, Artifical Intelligence, etc. You can find a list of all talks here.

One of the very interesting talks at Digital Xchange which attracted a large group of audience was "Deep Learning mit Keras und TensorFlow" given by Henning Buhl. He is a Computer Science Bachelor student at TH Köln who also works part-time under supervision of Prof. Wolfgang Konen. His talk coverd a large range of themes from neural networks to machine learning and Python code examples. As part of his talk he gave a live demo on Google Colaboratory with Jupyter notebooks showing TensorFlow and Keras examples. These examples are available on GitHub for your own study if you like. You can find them here or here on GitHub.

ECDA is an international conference with the main focus on data science and data mining which is being held in different European cities since 2013 almost every year. This year the conference was held in Paderborn, Germany. This 3-days conference hosted many interesting keynotes and a large number of talks given by scientists from all over the world. Prof. Wolfgang Konen organized and chaired the "Time Series Analysis and Online Algorithms"  session. In the same session Markus Thill presented his recent work "Online Adaptable Time Series Anomaly Detection with Discrete Wavelet Transforms and Multivariate Gaussian Distributions" and he received valuable remarks and feedback. The deadline for the paper  submission is on 15th of September and after successfully passing through a reviewing process the work will be published in Archives of Data Science, Series A Journal.

The R package TDMR (Tuned Data Mining in R) is now available on CRAN in a major new release 2.0. It supports the new R  package SPOT 2.0 (Sequential Parameter Optimization Toolbox) with its largely redesigned and simplified interface. TDMR 2.0 has as well a simplified interface. TDMR documentation and TDMR tutorials have been rewritten to account for the simpler interface.

Tuned Data Mining in R ('TDMR') performs the complete tuning of a data mining task (predictive analytics, that is classification and regression). Preprocessing parameters and modeling parameters can be tuned simultaneously. It incorporates a variety of tuners (among them 'SPOT' and CMA with package 'rCMA') and allows integration of additional tuners. Noise handling in the data mining optimization process is supported, see Koch et al. (2015) <doi:10.1016/j.asoc.2015.01.005>.

More information on TDMR is available on the TDMR project page.