A neural network approach for retailer risk assessment in the aftermarket industry

Advanced deep learning approaches to predict supply chain risks under COVID-19 restrictions

The ongoing COVID-19 pandemic has created an unprecedented predicament for global supply chains (SCs). Shipments of essential and life-saving products, ranging from pharmaceuticals, agriculture, and healthcare, to manufacturing, have been significantly impacted or delayed, making the global SCs vulnerable. A better understanding of the shipment risks can substantially reduce that nervousness. Thenceforth, this paper proposes a few Deep Learning (DL) approaches to mitigate shipment risks by predicting "if a shipment can be exported from one source to another", despite the restrictions imposed by the COVID-19 pandemic. The proposed DL methodologies have four main stages: data capturing, de-noising or pre-processing, feature extraction, and classification. The feature extraction stage depends on two main variants of DL models. The first variant involves three recurrent neural networks (RNN) structures (i.e., long short-term memory (LSTM), Bidirectional long short-term memory (BiLSTM), and gated recurrent unit (GRU)), and the second variant is the temporal convolutional network (TCN). In terms of the classification stage, six different classifiers are applied to test the entire methodology. These classifiers are SoftMax, random trees (RT), random forest (RF), k-nearest neighbor (KNN), artificial neural network (ANN), and support vector machine (SVM). The performance of the proposed DL models is evaluated based on an online dataset (taken as a case study). The numerical results show that one of the proposed models (i.e., TCN) is about 100% accurate in predicting the risk of shipment to a particular destination under COVID-19 restrictions. Unarguably, the aftermath of this work will help the decision-makers to predict supply chain risks proactively to increase the resiliency of the SCs.

Towards a Very Fast Feedforward Multilayer Neural Networks Training Algorithm

**This paper presents a novel fast algorithm for feedforward neural networks training. It is based on the Recursive Least Squares (RLS) method commonly used for designing adaptive filters. Besides, it utilizes two techniques of linear algebra, namely the orthogonal transformation method, called the Givens Rotations (GR), and the QR decomposition, creating the GQR (symbolically we write GR + QR = GQR) procedure for solving the normal equations in the weight update process. In this paper, a novel approach to the GQR algorithm is presented. The main idea revolves around reducing the computational cost of a single rotation by eliminating the square root calculation and reducing the number of multiplications. The proposed modification is based on the scaled version of the Givens rotations, denoted as SGQR. This modification is expected to bring a significant training time reduction comparing to the classic GQR algorithm. The paper begins with the introduction and the classic Givens rotation description. Then, the scaled rotation and its usage in the QR decomposition is discussed. The main section of the article presents the neural network training algorithm which utilizes scaled Givens rotations and QR decomposition in the weight update process. Next, the experiment results of the proposed algorithm are presented and discussed. The experiment utilizes several benchmarks combined with neural networks of various topologies. It is shown that the proposed algorithm outperforms several other commonly used methods, including well known Adam optimizer.

A Novel Fast Feedforward Neural Networks Training Algorithm

In this paper1 a new neural networks training algorithm is presented. The algorithm originates from the Recursive Least Squares (RLS) method commonly used in adaptive filtering. It uses the QR decomposition in conjunction with the Givens rotations for solving a normal equation - resulting from minimization of the loss function. An important parameter in neural networks is training time. Many commonly used algorithms require a big number of iterations in order to achieve a satisfactory outcome while other algorithms are effective only for small neural networks. The proposed solution is characterized by a very short convergence time compared to the well-known backpropagation method and its variants. The paper contains a complete mathematical derivation of the proposed algorithm. There are presented extensive simulation results using various benchmarks including function approximation, classification, encoder, and parity problems. Obtained results show the advantages of the featured algorithm which outperforms commonly used recent state-of-the-art neural networks training algorithms, including the Adam optimizer and the Nesterov’s accelerated gradient.

A six-dimensional model for supply chain sustainability risk analysis in telecommunication networks: a case study

Purpose

The main purpose of the present study was to improve and develop previously proposed models for Iran telecommunication networks. It should be noted that the six-dimensional (6D) sustainability model used in this study will be a useful and comprehensive model for industries. Since, the new dimension of IT along with the five well-known economic, social, environmental, technical and institutional aspects of organizations is considered to have great impacts on supply chain sustainability; the proposed framework can be practical.

Design/methodology/approach

According to the related literature review, there are two research streams in supply chain management. The first stream is exploratory research, seeking out conceptual discussions in this area. The second one is associated with mathematical models and techniques, aiming to set decision-making rules in this regard (Agrell et al., 2004). In this study, evaluation was performed using the FMEA method as an analytical technique based on the principle of pre-occurrence prevention to identify potential failure factors in sensitive systems (Mohammadfam and Kianfari, 2008).

Findings

After identifying the risks and causes of the incidence and effects and consequences of risks, preventive and risk control measures and advisory strategies were presented. Customers with 45.76% share in critical risks are threatening to maintain supply chain in these companies. During this study, it was found that 33.9% of the main source of supply chain critical risks was customers, constituting 45.76% of such risks, accompanied by organization, having a 38.88% share of critical risk generation. The study findings also revealed that 33.9% of critical risks were mainly (equally) related to economic and technical aspects of supply chain sustainability in telecommunication networks. Moreover, as a newly-introduced sixth dimension, IT represented 10.17% of critical risks threatening supply chain sustainability in such networks. Critical risks are mainly related to the economic and technical aspects (equally) with the sustainability of the telecommunication networks supply chain. Also, as a new finding and the sixth dimension, 10.17% of the critical risks that threaten the sustainability of the telecommunication networks supply chain have the information technology dimension.

Originality/value

The internet and fixed and mobile data services are provided by several private companies in Iran, which are relatively similar in terms of their supply chains. In order to manage the sustainability of Iran's telecommunication supply chain, telecommunication networks affiliated to Iran Telecommunication Company (ITC), operating in the field of data and internet services and fixed telephone were selected in 31 provinces. The intended networks were also providing an important part of the country's needs including Mobile Telecommunication Company of Iran, a subsidiary of Iran's telecommunication networks, as one of the top companies in this industry. Accordingly, all the networks studied in this study needed to be identified with regard to communication sustainability risks, since they provide management solutions to each other by segregating risks. In this study, 68 managers and 72 experts participated in different work teams of telecommunication networks.

V. An investigational analysis on forecasting intraday values

Purpose

The algorithmic trading has advanced exponentially and necessitates the evaluation of intraday stock market forecasting on the grounds that any stock market series are foreseen to follow the random walk hypothesis. The purpose of this paper is to forecast the intraday values of stock indices using data mining techniques and compare the techniques’ performance in different markets to accomplish the best results.

Design/methodology/approach

This study investigates the intraday values (every 60th-minute closing value) of four different markets (namely, UK, Australia, India and China) spanning from April 1, 2017 to March 31, 2018. The forecasting performance of multivariate adaptive regression spline (MARSplines), support vector regression (SVR), backpropagation neural network (BPNN) and autoregression (1) are compared using statistical measures. Robustness evaluation is done to check the performance of the models on the relative ratios of the data.

Findings

MARSplines produces better results than the compared models in forecasting every 60th minute of selected stocks and stock indices. Next to MARSplines, SVR outperforms neural network and autoregression (1) models. The MARSplines proved to be more robust than the other models.

Practical implications

Forecasting provides a substantial benchmark for companies, which entails long-run operations. Significant profit can be earned by successfully predicting the stock’s future price. The traders have to outperform the market using techniques. Policy makers need to estimate the future prices/trends in the stock market to identify the link between the financial instruments and monetary policy which gives higher insights about the mechanism of existing policy and to know the role of financial assets in many channels. Thus, this study expects that the proposed model can create significant profits for traders by more precisely forecasting the stock market.

Originality/value

This study contributes to the high-frequency forecasting literature using MARSplines, SVR and BPNN. Finding the most effective way of forecasting the stock market is imperative for traders and portfolio managers for investment decisions. This study reveals the changing levels of trends in investing and expectation of significant gains in a short time through intraday trading.

Scenario analysis of smart, sustainable supply chain on the basis of a fuzzy cognitive map

Purpose

This paper aims to present a smart, sustainable supply chain practices structure on the basis of the relational view.

Design/methodology/approach

A method based on fuzzy cognitive map was applied to construct a relational map to introduce and implement such relational methods. Considering this relational map as a guideline, observations into particular methods and ways of applying relational methods to attain sustainable development goals across organizations has been introduced.

Findings

Primary outcomes provided a series of relational methods for the purpose of giving advice to those organizations and their suppliers for smart, sustainable supply chain. Reliance between relational methods were examined and assessed under seven meaningful groups: economic internet of things (IoT), green internet of things, social internet of things, economic supply chain, green supply chain, social supply chain and other variables.

Practical implications

This study guides managers toward an improved perception of the connection among IoT instances and sustainable supply to modeling smart, sustainable supply chain. Managers can determine the practices that need more focus along with the practices that are less relevant. Thus, this will help managers in the decision-making process and to organize their decisions by planning and calculating the relative importance and influence of smart, sustainable practices on each other and on the company’s smart, sustainable program.

Originality/value

To the best of the authors’ knowledge, this is the first approach that promptly examines and determines the interdependencies between relational methods and constructs a relational map for the purpose to introduce and analyze smart, sustainable supply chain.