Prof. Defu Zhang, Xiamen University, China
Research Area: Big data, Business intelligence, Knowledge graph, Operation research
Title：Big data platform for agriculture supply chain
Abstract：
There are many problems such as Blind planting, planting technology and precision marketing and brand building, we use big data and artificial intelligence technology to solve these problem. A big data platform with production forecasting system, agriculture expert system with Pest identification, precision marketing system and Food safety traceability system is developed. This platform has been used to help the farmers sell agriculture products
Assoc. Prof. Desislava Stoilova, South-West University “Neofit Rilski”, Bulgaria
Research Area: Public finance, Local finance, Financial management, Economic growth
Title：THE IMPACT OF THE EU FUNDS ON REGIONAL ECONOMIC DEVELOPMENT: EVIDENCE FROM BULGARIA
Abstract：
The purpose of this paper is to study the impact of the EU funds on regional economic development in Bulgaria for the period 2007 – 2018. The descriptive analysis is focused on the general trends in economic development of the six NUTS 2 regions (Southwest, South Central, Southeast, Northwest, North Central, and Northeast). The results show that in the last decade socio-economic development in Bulgaria demonstrated very strong territorial concentration. The main engine of economic growth is the Southwest region, which produces nearly half (48.2%) of the gross domestic product of Bulgaria, while the differences between the other regions are significantly smaller (NSI, 2020). The main factor for the faster development of the Southwest region is the capital city Sofia, which concentrates a significant part of the national economy.
Assoc. Prof. Carol Hargreaves, National University of Singapore, Singapore
Research Area: Artificial Intelligence, Machine Learning & Deep Learning
Title：Big Data Application Of Deep Learning In Forecasting Stock Prices
Abstract：
Sequence prediction problems are considered as one of the hardest problems to solve in the data science industry since sequence prediction problems have to deal with autocorrelation, volatility clustering,
non-Gaussianity, and possibly cyclical data.
Deep Learning (DL) and Reinforcement Neural Networks (RNNs) in particular, can deal with autocorrelation, volatility clustering, non-Gaussianity and cyclical data and can help to make accurate predictions under these conditions. We propose to use a Long Short-Term Memory (LSTM) network, similar to the RNN but where the hidden units are replaced by memory cells to predict the stock returns for stocks from the Australian Stock Market.
For stock we predict the return at time t+1 using historical returns up to time. The predictions from the LSTM model help us to decide at time t which stocks to buy, hold or sell. This way we have an automated trading policy. The LSTM network will be trained for periods of three years (2016 – 2018), validated on (2019) followed by testing on future data.
This study intends to show that the deep learning LSTM network can help us pick good profitable portfolios.
