Si avvisano gli studenti che lunedì 14 novembre alle ore 15:00 in Aula NAI si svolgerà il seminario dal titolo “Intelligent forecasting. Predict time series forecasting enhanced with external signals”, svolto dal dott. Marocchino di KPMG.

Gli studenti che hanno in calendario la lezione con il prof. Vassallo si recheranno in Aula NAI, dove troveranno lui e gli altri colleghi.

 

Dato il tema interessante e l’opportunità di entrare in contatto con la realtà della Data Science in una delle più grandi aziende di consulenza, la partecipazione è fortemente consigliata.

 

Di seguito l’abstract del seminario:

 

Time-series forecasting is today the backbone of many engineering and economic processes, supporting decision making. Artificial Intelligence and specifically Machine Learning has given a fundamental boost with non-stationary time-serie forecasting models. While forecasting of time-series on their own are of some relevance (we can refer to this prediction as historical-seasonal prediction), external effects (exogenous signals) might influence the time-series evolution and so the results, i.e. the pandemic in many aspects of our daily life and in the business sector. In this seminar we will present the data lifecycle tailored on the KPMG forecasting suite, how from the KPMG Signals Repository composed of more than 60 000 signals we down select the key exogenous ones that effectively contribute to the prediction. We will also offer a focus on some implemented algorithms, e.g. LSTM algorithms.