UJ’s Prof Ilse Botha explores how time series forecasting evolved in the wake of Big Data and AI

​Emerging technologies could lead to the next quantum leap in how data is collected; how data is analysed; and how analysis is used for policymaking and the achievement of better forecasts. Ilse Botha, a Professor in the Department of Accountancy at the University of Johannesburg (UJ), argued that the advanced skills set required to work with Big Data present a major challenge since there is a shortage of data scientists.

Researchers are highly proficient in using traditional statistical techniques to obtain accurate forecasts, but the incompatibility of traditional statistical techniques with unstructured and large data sets are impeding the effectiveness and application of forecasts from Big Data. Prof Botha believes it is important that curricula is developed to incorporate the skills necessary to analyse Big Data in order for future statisticians or econometricians to be well equipped with the required skills when she delivered her inaugural address in the Council Chambers, Madibeng Building, Auckland Park Kingsway Campus on Tuesday, 23 July 2019.

Prof Botha’s address focused on the evolution and future of time series analysis and forecasting through the lens of artificial intelligence (AI). The theme was entitled ‘Time Series Forecasting in the Artificial Intelligence domain: Learning through the lens of time’

“The challenges that researchers face are how to build time series models that significantly learn to forecast from Big Data sets; and, how to improve forecasting with learning models when we have limited observations. It is also important to determine the implications for building forecasting systems that can handle large data volumes. Subjects such as the black box in learning models, confidence intervals and uncertainty of forecasts should be considered.”

“Due to this data revolution, challenges and opportunities exist for research, education, current research fields and new research fields. A new skill set is necessary due to the availability of data in certain fields, and the advances in computational power, technology and deep learning (DL) empower researchers to handle bigger time series data sets with better speed and accuracy. Various approaches give the flexibility to choose the best method according to the characteristics of the data,” said Prof Botha.

Prof Botha suggested that to overcome the constraints imposed by skills in other disciplines, modules should also be incorporated in various programmes to develop skills required to understand and analyse Big Data, using novel techniques. “The application of Big Data and time series models in accounting and auditing is still in its infancy, but due to the availability of Big Data the opportunity is promising.”

She added that quality financial and nonfinancial information is very important and this requires the use of technology to improve financial reporting and audit processes. Business analytics, along with access to detailed industry information, aids businesses in identifying challenges and opportunities that can create business value,” explained Prof Botha. “Time series forecasting is central to the automation and optimisation of business processes such as supply chain, cloud computing and workforce scheduling.”


prof ilse botha
Prof Ilse Botha
Share this

Latest News

All News