Machine learning for supply chain's big data: State of the art and application to social networks' data

Abstract

In the context of today 's pattern of globalization and a huge amount of information, a smart supply management chain is required. Naturally, statistics and operations research are used for optimizing supply and demand objectives. However, the new context brings out new opportunities at descriptive, predictive and prescriptive levels for supply chain network design, logistics and distribution and strategic sourcing. The key question is still how to capture and to use information. One striking example can be taken from social media, where their use allow to gain insight into the perception of consumers and to capture a real time overview of consumer reactions, regarding one or more specific events. In this regard, different modern approaches, such as IoT or Quantum neural network, are developed. In the same line of thought, we propose an analytic approach, based on KNN, Logistic Regression and SVM with the use of Twitter data in chicken supply chain management. Results identify the main concerns related to chicken products and allow to the development of a consumer-centric supply chain. The proposed approach can be extended to other topics such as anomaly detection and codification of customer intelligence.