Master Thesis: Accounts Receivable Cash Flow Forecasting

In this paper I investigate to what extent machine learning techniques can predict the cash flow of the accounts receivable based on historical data at a company of which the accounts receivable consists of few large value entries and many small value entries. In this study I compare the performance of SES, ARIMA, LSTM and Hybrid ARIMA-LSTM models on one company dataset and three synthetic datasets. The python code used in the models can be found within the paper where it is also explained how I optimized each model.