The article explores different models for predicting interest rates and valuing financial products. It looks at short-term interest rate models and no-arbitrage term structure models. Short-term models try to match real data closely, while no-arbitrage models aim to match market prices. The study finds that parameter estimates in these models can vary a lot depending on the data and methods used. This means the models should be used carefully. The researchers also consider how non-linearities in interest rate changes can affect predictions.