Let's discuss the need for modeling a hedge fund. Mr. Jefferson has just switched over to a niche hedge fund. He has years of experience in handling larger portfolios, but he has decided to shift to a smaller hedge fund as a partner. The place isn't as organized as the larger hedge funds with access to proprietary algorithms that run based on machine learning principles and manage a lot of the buy and sell transactions based on the hedge fund's goals. In this smaller and newer hedge fund, the aim is to slowly build a portfolio that delivers market-leading returns, thereby attracting newer clients. Mr. Jefferson has been asked, by the director, to recommend stocks that he would like to invest in, and his initial first-month portfolio has been assigned £10 million.

One of the stocks that Jefferson noticed that the hedge fund doesn't invest its client money in is a particular mobile phone manufacturer. It seemed peculiar to him that the fund has resisted the temptation to invest in a stock that most clients would like to hold.

To convince his director to invest in this stock, Jefferson needed to write a research report describing various aspects of the stock. Apart from doing the technical analysis, qualitative research, and other tasks for the report, he wanted to build a forecasting model. The research report would probably convince the director to invest. But the forecasting model would showcase that the hedge fund could predict the stock movement with a certain degree of accuracy. This would enable the hedge fund to actively manage the investments. Jefferson's goal wasn't just to invest his portfolio allocation; he wanted to create a proprietary algorithm for managing funds at the hedge fund. The forecasting model was just a start, to get the conversation going with the director.

Jefferson summarized the business problem:

• Build a forecasting model to predict the stock of a mobile phone manufacturer
• Assess its predicting power