Text generation for financial reports
- 1 min The master thesis gives an insight into the theory and application of different methods of text generation. Since many areas of the economy are related to text-based problems, this thesis, in order to learn the financial language, used a data source of the financial statements of more than 5000 companies. With the help of different algorithms based on neural networks, an attempt was made to automatically generate text based on a given topic. This automation of recurring and repeatable tasks could save time and avoid human error in many situations. The master thesis is mainly addressed to people with a strong background in statistical and mathematical methods. The theory and practical application of Deep Learning is not a prerequisite, as the basic concepts are described extensively in the thesis.
Since I had some problems to combine the information from the different sources I created the graphic to the left which shows a combination of the blog post and the post here where I think the graphics are very helpful but an error in explaining the number_units is present. Check the Stack Overflow Answer for detailed information and the full size graphic.