13 April 2017 – A team of NUS Computing students, comprising Year 4 Business Analytics students Che Mingzhou and Wu Zijing, and Year 4 Information Systems student Li Dongyan, won the second prize at the inaugural Institute for Operations Research and the Management Sciences (INFORMS) Operations Research and Analytics Student Team Competition.
Of the many teams from around the world that submitted entries, eight were eventually selected to present their findings at the finals which were held at the INFORMS Business Analytics Conference, in early April, in Las Vegas, USA. Sponsored by Syngenta, a global Swiss biotechnology company that produces crop chemicals and seeds, the competition challenged teams to develop a model to predict the sales potential of soybean seed varieties and select the best-performing ones for commercial release, based on historical data on the varieties’ experiment results and sales.
Speaking on behalf of the team, Mingzhou said that data crunching was the hardest part of the competition because it tested skills from multiple disciplines such as business, statistics and engineering. However, he believed that his team stood out because of their creativity, multi-disciplinary background and business orientation. “Round after round of discussion and data crunching, we created features that few teams had thought of. Also, we combined traditional statistical approaches and machine learning models to produce a more comprehensive and accurate solution. Lastly, during the presentation, instead of mentioning too many technical details, we focused more on business insights and kept things straightforward,” Mingzhou said.
The team credited their education at NUS Computing for their performance in the competition. “The Business Analytics programme covers most of the knowledge required to complete, if not excel in, a data analytics competition [like this]. The Information Systems [programme] also provides training in presentation and report-writing, as well as basic understanding of both business and analytics,” Mingzhou said.
Describing her experience of the competition, teammate Zijing said, “It is great to know that analytics is being acknowledged and applied in more and more areas, including more traditional industries like agriculture. Having witnessed people's passion during the competition and the conference, I have more confidence in the potential of analytics, especially applied in different aspects of operational research.” Adding to that, Dongyan said that the distinct approach her team had taken made her realise the importance of understanding what a company needs. Referring to the competition’s task of developing a model, she elaborated, “Some teams prioritised prediction accuracy, while our goal was to achieve a balance between prediction accuracy and model interpretability. Ideally, a company should strive to achieve the most accurate model. However, such a model may require the company to have an IT system to implement it, which may be costly. In the short-term, a model that is not the most accurate but can generate rules that are easy to understand and simple to apply, may be a more suitable solution to the company.”
For coming in second, the team was awarded a cash prize of US$5000. All eight teams also received a $2,500 stipend to offset the cost of conference registration and travel.