I have three full years of trend data sitting in a single Excel workbook. Each row records a two-digit value, and I now need a machine-learning model that can reliably forecast both the “inside” and “outside” four-digit combinations drawn from the digits 0-9. Historical trends are the only signal I want the model to learn from; please do not add external data or random sampling. Target performance is 90 % prediction accuracy, measured on a hold-out set we will agree on before training begins. I will share the Excel file once we start; you deliver: • Clean, well-commented code (Python preferred) that loads the spreadsheet, prepares the data, trains the model, and outputs predictions. • A brief write-up of feature engineering, model choice, and validation results demonstrating the required accuracy. • A simple way for me to paste new two-digit data and receive fresh four-digit predictions—command-line script, Jupyter notebook, or lightweight GUI is fine. Please outline the algorithms or libraries you plan to use (e.g., pandas, scikit-learn, TensorFlow, PyTorch) and how you will ensure the accuracy threshold is met. I am ready to start as soon as I confirm the approach fits the constraints above.