Sahar Behzadi

In cooperation with the research group "Data mining in medicine" at LMU, we have worked on clustering of mixed-type data (i.e., a mixture of categorical and numerical attributes).

My focus has been on mining heterogeneous data sets. In particular, I worked on clustering of mixed-type data (categorical and numerical attributes), causal inference, and anomaly detection of heterogeneous time series.

I have taught the following courses at the university: Advanced Topics in Data Analysis, Data Mining, Foundation of Data Analysis, Scientific Data Management, and Mathematics.

In the research group "Machine Intelligence", we investigated clustering methods to jointly learn the parameters of a neural network and the cluster assignments of the resulting features. Later, we applied this approach to the unsupervised training of convolutional neural networks on large data sets including images from additive manufacturing. We employed Tensorflow to implement the approach and develop a library.

In the research group " Integrative Knowledge Discovery in Databases (iKDD)", we have developed non--parametric clustering algorithms for mixed-type data sets.

In this company, our focus has been on Hospital Information Systems (HIS) and Management Information Systems (MIS).

My focus has been on mining heterogeneous data sets. In particular, I worked on clustering of mixed-type data (categorical and numerical attributes), causal inference, and anomaly detection of heterogeneous time series.