I have a clean dataset of around 200,000 observations over the last 15 years of certain financial events, with around around 10 features that can be used to predict returns. I have attached an ipython notebook with my work so far, and I'm looking for an expert to help me take this further.
Note that because this is event data, instead of the normal random split of the dataset into training / testing, we need to only use historical data in the training set.
If this works out well, there are many other similar datasets that I wish to analyze, and this will be a long term relationship with challenging, interesting and well compensated work.
Skills: Machine learning (PCA, ANN, SVM, feature scaling, parameter optimization, ensemble algorithms, algorithm evaluation techniques), time series analysis, data visualization
Packages: Python data analysis (ipython notebook, pandas, sklearn, seaborn / matplotlib)
Must have strong verbal and written communication in English.