Fashion retail forecasting by evolutionary neural networks

Recent literature on nonlinear models has shown that neural networks are versatile tools for forecasting. However, the search for an ideal network structure is a complex task. Evolutionary computation is a promising global search approach for feature and model selection. In this paper, an evolutionary computation approach is proposed in searching for the ideal network structure for a forecasting system. Two years’ apparel sales data are used in the analysis. The optimized neural networks structure for the forecasting of apparel sales is developed. The performances of the models are compared with the basic fully connected neural networks and the traditional forecasting models. We find that the proposed algorithms are useful for fashion retail forecasting, and the performance of it is better than the traditional SARIMA model for products with features of low demand uncertainty and weak seasonal trends. It is applicable for fashion retailers to produce short-term retail forecasting for apparels, which share these features.     

Investigating Adoption of Free Beta Applications in a Platform‐Based Business Ecosystem

Planning new product development (NPD) activities is becoming increasingly difficult, as contemporary businesses compete at the level of business ecosystems in addition to the firm‐level product‐market competition. These business ecosystems are built around platforms interlinking suppliers, complementors, distributors, developers, etc. together. The competitiveness of these ecosystems relies on members utilizing the shared platform for their own performance improvement, especially in terms of developing new valuable offerings for end users. Therefore, managing the development of the platform‐based applications and gaining timely end‐user input for NPD are of vital importance both to the ecosystem as a whole and to the developers. Subsequently, to succeed in NPD planning developers utilizing beta testing need a thorough understanding of the adoption dynamics of beta products. Developers need to plan for example resource allocation; development costs; and timing of commercial, end‐product launches. Therefore, the anticipation of the adoption dynamics of beta products emerges as an important antecedent in planning NPD activities when beta testing is used for gaining end‐user input to the NPD process. Consequently, we investigate how free beta software products that are built upon software platforms diffuse among their end users in a cocreation community. We specifically study whether the adoption of these beta products follows Bass or Gompertz model dynamics used in the previous literature when modeling the adoption of stand‐alone products. Further, we also investigate the forecasting abilities of these two models. Our results show that the adoption dynamics of free beta products in a cocreation community follow Gompertz's model rather than the Bass model. Additionally, we find that the Gompertz model performs better than the Bass model in forecasting both short and long out‐of‐sample time periods. We further discuss the managerial and research implications of our study.     

WHAT HAPPENS WHEN DEMAND IS ESTIMATED WITH A MISSPECIFIED MODEL?*

We conduct Monte Carlo experiments to investigate the biases of assuming a misspecified demand model. We study continuous models (linear, log‐linear and AIDS), and discrete choice models (logit) in the context of differentiated products and aggregate data. Estimating demand with the ‘wrong’ model yields varying degrees of bias in estimated elasticities, but the logit model can yield unbiased estimates for a certain size of the assumed market potential. Merger simulations confirm the key importance of market potential in logit estimation suggesting that a discrete choice model may be preferable even when the discreteness of the purchase decision is questionable.     

Robust analogizing and the outside view: two empirical tests of case‐based decision making

Executives use analogies to improve strategic decisions. However, existing research provides little guidance on the types and number of analogies that produce the best decisions. We examine models of analogy and present findings from two empirical tests. The first test, a study of private‐equity investment decisions, finds that an ‘outside view’—forming a reference class of analogies—performs better than a few analogies familiar to the decision maker. The second test, a study of film revenue forecasts, compares a new model we call similarity‐based forecasting (SBF) with existing methods. The study finds that SBF, which combines elements of reference class forecasting and case‐based decision making, produces better forecasts than regression models. We discuss the consequences of our findings for research and practice. Copyright © 2012 John Wiley & Sons, Ltd.     

Neural network application for fuzzy multi-criteria decision making problems

In this paper, a fuzzy multi-criteria decision making model is presented based on a feed forward artificial neural network. This model is used to capture and represent the decision makers' preferences. The topology of the neural network model is developed to train the model. The proposed model can use historical data and update the database information for alternatives over time for future decisions. Basically, multi-criteria decision making problems are formulated, and neural network is used to learn the relation among criteria and alternatives and rank the alternatives. We do not use any utility function for the modeling; however, a unique method is proposed for eliciting the information from decision makers. The proposed model is applicable for a wide variety of multi-attribute decision making problems and can be used for future ranking or selection without managers' judgment effort. Simulation of the managers' decisions is demonstrated in detail and the design and implementation of the model are illustrated by a case study.     

COST PASS‐THROUGH IN DIFFERENTIATED PRODUCT MARKETS: THE CASE OF U.S. PROCESSED CHEESE*

In this paper, we estimate a mixed logit model for demand in the U.S. processed cheese market. The estimates are used to determine pass‐through rates of cost changes under different behavioral regimes. We find that, under collusion, the pass‐through rates for all brands fall between 21% and 31% while, under Nash‐Bertrand price competition, the range of pass‐through rates is between 73% and 103%. The mixed logit model provides a more flexible framework for studying pass‐through rates than the logit model since the curvature of the demand functions depends upon the empirical distribution of consumer types.     

Transitions Out of Casual Employment: The Australian Experience

This paper uses longitudinal data to examine the extent to which casual employees, who account for almost 25 percent of all Australian employees, are able to access non‐casual jobs in the future, and to contrast their experiences with that of other labor market participants. A dynamic mixed multinomial logit model of labor market states is estimated which reveals high rates of mobility from casual employment into non‐casual employment. Among men, casual employees are found to be far more likely to make the transition into non‐casual employment than otherwise comparable unemployed job seekers. For women, however, this is not the case.     

Exploring geographic variation in corporate broadband adoption; evidence from Irish small- and medium-sized enterprises (SMEs)

Despite its growth experience, Ireland still lags behind in international league tables on broadband adoption and within Ireland great disparities exist between central and peripheral regions. This paper explores the context-specific determinants of broadband adoption among Irish small- and medium-sized enterprises (SMEs), attempting to shed some light on the sources of the considerable geographic variation in particular. Drawing from determinants of broadband adoption identified in literature the authors develop a number of plausible hypotheses. Using cross-section data from a ComReg survey of Irish-based SMEs, a logit model of broadband adoption is estimated. Among other factors, a company's industrial sector and other demand proxies are good predictors of broadband adoption. Controlling for other factors, regional ISP market concentration appears to be negatively associated with the probability of broadband adoption. It is argued that, in the absence of more detailed information, statistics on regional-level market structure could be a promising indicator of the supply-side.     

支持向量机模型在断层破碎带围岩变形预测中的应用

围岩变形预测是隧道安全评价及其指导后期施工的重要依据,为提高变形预测精度,结合工程实践,提出了PSO-SVM-BP预测模型的思路。首先,利用三次样条插值及二次平滑法对变形数据进行预处理,为后期变形预测奠定基础;其次,利用粒子群算法对支持向量机进行参数优化,建立PSO-SVM模型,并对围岩变形进行初步预测;最后,利用BP神经网络进行误差修正,达到综合预测的目的,并利用工程实例进行检验,以验证预测模型的有效性。结果表明:初步预测结果的相对误差均小于5%,而误差修正后的预测精度被提高到0.97%,预测精度较高,验证了预测模型的有效性,可为类似研究提供参考。     

The Determinants of Technology Adoption in Italian Manufacturing Industries

The adoption of new technologies in Italian manufacturing industries is analysed using data for 13,334 firms selected from the 1990–92 Community Innovation Survey. The determinants of technology adoption are analysed in an econometric framework (logit model) which is a general test of different theoretical explanations of technological diffusion. We particularly refer to the rank, epidemic and information effects which significantly affect the use of new technology in Italian manufacturing industries. We use a set of explanatory variables which enables us to set up a well specified empirical model and to use odds ratios to determine the effect of their changes on the adoption probability, thus giving a more precise picture of the determinants of technology adoption.     

The adoption of radical manufacturing technologies and firm survival

The emergence of dramatically innovative, or radical, new manufacturing technologies can force pivotal and life‐threatening decisions for industry competitors. These technologies can represent a huge cost for adopting firms, but may also offer the chance to achieve competitive advantage through superior manufacturing. While prior research has considered a range of production process decisions (e.g., JIT, mass customization) and outcomes for end‐product technologies, little attention has been given to adoption decisions relative to core manufacturing technologies. This study examines an industry's adoption of major manufacturing technologies over several decades and demonstrates that two groups of contingencies related to adoption (e.g., timing and cumulative effects) have a significant impact on firm performance. Based on a sample of over 1,000 firms, the results provide insights into the effects of adoption timing and ‘manufacturing technology bundles’ on firm survival. We also find that adoption of manufacturing technologies prior to the inflection point of the estimated Bass diffusion curve for each technology leads to significant reduction in firm mortality. Thus, we are able to demonstrate the ability of the Bass model to predict the survival outcomes of firms facing manufacturing technology adoption decisions. The strategic implications of these pivotal decisions are considered. Copyright © 2008 John Wiley & Sons, Ltd.     

SPLINTERING AND INERTIA IN NETWORK INDUSTRIES*

We analyse a stylised game of technology adoption with network effects and two new technologies. Potential adopters can adopt early, late or not at all. We show that one of the reasons for the failure of new technologies can be the presence of multiple incompatible variants of that technology. An adopter's individual incentives to adopt are lower with two technologies than with one. Turning to aggregate expected welfare, we find that two active technologies may be welfare‐improving.     

Postponed product differentiation with demand information update

This paper studies the problem of how to coordinate postponed product differentiation and forecast update to improve manufacturing efficiency. We consider a two-stage model of multiple products with a common component. In stage 1, the manager obtains a prior demand distribution of each product and decides the production quantity of the common component. In stage 2, the demand forecast is updated and the common component is differentiated into various final products. Then the final demand of each product is realized and inventory leftover (shortage) is assessed. We use stochastic programming to model this problem, and propose an optimal bundle-type algorithm to solve it. Furthermore, we develop some simple and effective approximation algorithms for several special cases. Extensive numerical experiments are conducted to show the effectiveness of the approximation algorithms, to compare the performance between the traditional production model and the postponement production model, and to examine the impact of parameters on the performances of the two systems.     

Technology Adoption in Online Social Networks

With the advances of information technology, online social networks are becoming increasingly important venues for technology adoption. However, although the dynamics of technology adoption in real world social networks have been well documented, technology adoption in online social networks remains relatively under‐explored. This study identifies the differences between online and offline social networks and proposes a framework to investigate the dynamics of technology adoption in online social networks. To illustrate the proposed research framework, this study employs behavior‐link panel data obtained from an open source software (OSS) development network to examine how online social networks affect the adoption of Subversion, the latest OSS version control technology. Based on social network theory, co‐membership is used to construct online social networks within the OSS development network. Methodologically, this study takes advantage of the panel dataset and addresses the issues of simultaneity and individual heterogeneity that frequently confound the relationship between network structure and adoption decision, and as a result it demonstrates a more compelling relationship between social networks and technology adoption. The results of this study reveal that social networks are major conduits for technology adoption in an online social network in terms of imitation, leadership, lock‐in, similarity, recency, and team size effects. In online social networks, one's decision to adopt a new technology is strongly influenced by the actions of the connected others. Project leaders have a stronger influence over other members in technology adoption decision making, even in informal virtual teams where traditional governance structures do not apply. Older projects exhibit stronger inertia and thus lack innovativeness. Similarities among projects facilitate faster adoption, and the effect of leadership attenuates in the networks with increasing project dissimilarity. Recent adoptions of technology within the networks, rather than more distant ones, have a stronger impact on subsequent adoption, implying the salience of memory over usage confidence, and increased size of a project team accelerates the rate of adoption. These results help in understanding the dynamics of technology adoption in online social networks, and provide useful guidelines for firms to promote technology and product innovation.     

Modelling and forecasting national introduction times for successive generations of mobile telephony

An accurate prediction of the timing of a country's introduction of a new generation of mobile telephony benefits numerous agents including suppliers of network and consumer equipment, regulators, and network planners. We consider the estimation and prediction of the time interval between the international introduction of a generation of mobile telephony and its introduction into a specific country when a decision maker judges the introduction of a newer technology a worthwhile investment. Using literature-based socio-economic and geographical variables, we examine how well variation in international introduction times of four generations of mobile telephony in 172 countries can be explained and forecast. We model and forecast introduction times at two levels of granularity: we use Cox's proportional hazards model for the introduction time; we partition countries into introduction time-based segments and model segment membership using multinomial logistic regression. Our modelling of each generation considers three subsets of explanatory variables: All variables, socio-economic Covariates only, Regional dummies only. Over successive generations, the Covariates only models reveal the changing relevance of each socio-economic covariate. Model-based forecasting of the introduction time of the next generation is performed under three hypotheses making different uses of the information available at the time the relevant generation is launched internationally. However, changing socio-economic environments coupled with changing models impair forecasting accuracy, the lower accuracy of modelled introduction times is concentrated in 20% of countries. We speculate about the nature of the unobserved factors affecting these countries' decision processes.     

Policy orientations and technology choices in standards wars

Standard-setting organizations (SSOs) exhibit a variety of policy orientations toward the conflicting interests of technology developers and adopters. In this paper, we analyze a model that incorporates the technology choices of SSOs in standards wars and royalty determinations made by the developers of essential technologies. We show that both policy orientations toward developers relative to adopters and coordinated standard setting by SSOs that issue competing standards may result in a more-than-optimal number of essential technologies. Furthermore, we examine how SSOs’ technology choices may be affected by both network effects in standard adoption and coalition formation among developers.     

Minimum Cash Wages,Tipped Restaurant Workers,and Poverty

This is the first study to examine the effect of increases in the tipped minimum cash wage—the wage employers must pay to tipped employees—on poverty. Using March Current Population Survey data (1988–2014), we find that tipped minimum cash wage increases are associated with declines in the risk of a tipped restaurant worker living in a poor family (elasticities around –0.2). However, we find little evidence of poverty‐alleviating effects when using the household rather than the family as the sharing unit. This result is consistent with evidence that a substantial share of tipped workers who live in a poor family live in a nonpoor household with persons unrelated by blood, marriage, or adoption who contribute to the household's income. Furthermore, we find that tipped minimum cash wage hikes are associated with increases in the risk of a younger, less‐educated individual living in a poor family or household. Adverse labor demand effects that redistribute income among low‐skilled individuals drive these results. We conclude that raising the tipped minimum cash wage is a poorly targeted policy to deliver income to poor restaurant workers.     

ECONOMIC APPRAISAL OF TRANSPORT PROJECTS: A Manual With Case Studies by Hans A. Adler

A challenge to developing data-driven approaches in finance and trading is the limited availability of data because periods of instability, such as during financial market crises, are relatively rare. This study applies a stability-oriented approach (SOA) based on statistical tests to compare data for the current period to a past set of data for a stable period, providing higher reliability due to a more abundant source of data. Based on an SOA, this study uses an artificial neural network (ANN), which is one of the commonly applied machine learning algorithms, for simultaneously forecasting the volatility and classifying the level of market stability. In addition, this study develops a dynamic target volatility strategy for asset allocation using an ANN to enhance the ability of a target volatility strategy that is established for automatically allocating capital between a risky asset and a risk-free cash position. In order to examine the impact of the proposed strategy, the results are compared to the buy-and-hold strategy, the static asset allocation strategy, and the conventional target volatility strategy using different volatility forecasting methodologies. An empirical case study of the proposed strategy is simulated in both the Korean and U.S. stock markets.     

Switching costs in the US seed industry: Technology adoption and welfare impacts

We evaluate the role of brand and technology switching costs in the US soybean seed industry using a unique dataset of actual seed purchases by about 28,000 farmers from 1996 to 2016. Using a random coefficients logit model of demand, we estimate brand and technology switching costs, characterize the distributions of buyers’ willingness to pay for seed brands and the glyphosate tolerance (GT) trait, and assess the implications of brand and technology switching costs for farmers’ welfare, technology adoption, firm profits, and firm market shares. We find that farmers are willing to pay large premiums for brand labels, and even larger premiums for the GT trait, although there is considerable heterogeneity in these values. Switching costs play an important role in the soybean seed industry. Eliminating these costs would significantly increase buyers’ welfare, reduce seed prices and firm profits, decrease adoption of the GT trait, and impact industry consolidation by expanding smaller firms’ market shares.     

The maximum capture per unit cost location problem

This paper considers the problem of siting p new facilities of an entering firm to a competitive market so as to maximize the market share captured from competitors per unit cost. We first formulate the problem as a mixed 0-1 fractional programming model, in which we incorporate the fixed cost and transportation cost. The model can deal with the case where some demand nodes have two or more possible closest servers. We then re-formulate the problem as a 0-1 mixed integer linear program. We use a one-opt heuristic algorithm based on the Teitz-Bart method to obtain feasible solutions and compare them with the optimal solutions obtained by a branch-and-bound algorithm. We conduct computational experiments to evaluate the two algorithms. The results show that both algorithms can solve the model efficiently and the model is integer-friendly. We discuss other computational results and provide managerial insights.