A comparison of several adaptive forecasting procedures

There is a considerable amount of published research relevant to the selection of automatic, online, single exponential smoothing models, applicable primarily to following a stationary process of data. Research into adaptive, double exponential smoothing models which are applicable to following a linear trend in the data has not been so profuse, however, particularly under strict online design specifications. Five feasible methods from the double exponential family were selected for comparison in this study, which consisted of four separate simulation experiments. The online design specifications are interpreted quite strictly to include adaptiveness, recursiveness, fading memory, computational economy, robustness, self-initialisation and comprehensibility to management.     

A further appreciation of demand forecasting models: Some methods based on survey information

This paper deals with the data available from survey sources which primarily provides cross-sectional statistics that can be applied in forecasting models. It is intended as a complement to an earlier paper by the authors which appeared in the September 1973 edition of Long Range Planning.     

The efficacy of using judgmental versus quantitative forecasting methods in practice

In an era where forecasts drive entire supply chains forecasting is seen as an increasingly critical organizational capability. However, business forecasting continues to rely on judgmental methods despite large advancements in information technology and quantitative method capability, prompting calls for research to help understand the reasons behind this practice. Our study is designed to contribute to this knowledge by profiling differences between firms identified as primary users of either judgmental or quantitative forecasting methods. Relying on survey data from 240 firms we statistically analyzed differences between these categories of users based on a range of organizational and forecasting issues. Our study finds large differences in forecast error rates between the two groups, with users of quantitative methods significantly outperforming users of judgmental methods. The former are found to be equally prevalent regardless of industry, firm size, and product positioning strategy, documenting the benefits of quantitative method use in a variety of settings. By contrast, the latter are found to have significantly lower access to quantifiable data and to use information and technology to a lesser degree.     

Integrating management judgment and statistical methods to improve short-term forecasts

The complementary strengths that management judgment and statistical methods can bring to the forecasting process have been widely discussed. This paper reviews research on the effectiveness of methods that are designed to allow judgment and statistical methods to be integrated when short-term point forecasts are required. The application of both voluntary and mechanical integration methods are considered and conditions identified where the use of particular methods is appropriate, according to current research. While acknowledging the power of mechanical integration methods that exclude the judgmental forecaster from the integration process, the paper suggests that future research effort should focus on the design of forecasting support systems that facilitate voluntary integration.     

A comparison between the VERT program and other methods of project duration estimation

Several authors have compared and proposed exact, asymptotic or simulation methods for estimating or deriving the duration of project networks. These authors have all concentrated on one aspect of uncertainty—time. The results of simulations obtained through the Venture Evaluation and Review Technique (VERT) are compared with those of other authors. A brief exposé of the extra facilities within VERT is also given. That is, the ability to jointly manipulate time, cost and performance measures, as well being able to specify the distribution from which to sample data upon these uncertain measures.     

A comparison of three weight elicitation methods: good, better, and best

This paper compares the properties and performance of three weight elicitation methods. It is in effect a “second round contest” in which the Bottomley et al. (2000) champion, direct rating (DR), locks horns with two new challengers. People using DR rate each attribute in turn on a scale of 0–100, whilst people using Max100 first assign to the most important attribute(s) a rating of 100, and then rate the other attributes relative to it/them. People using Min10 first assign the least important attribute(s) a rating of 10, and then rate the other attributes relative to it/them.The weights produced by Max100 were somewhat more test–retest reliable than DR. Both methods were considerably more reliable than Min10. Using people's test–retest data as attribute weights on simulated alternative values in a multi-attribute choice scenario, the same alternative would be chosen on 91% of occasions using Max100, 87% of occasions using DR, but only 75% of occasions using Min10. Moreover, the three methods are shown to have very distinct “signatures”, that is profiles relating weights to rank position. Finally, people actually preferred using Max100 and DR rather than Min10, an important pragmatic consideration.     

A model for forecasting energy economic systems

This paper presents a small integrated model which can be used for forecasting the behaviour of energy economic systems. The model is small, particularly when compared with the huge econometric modelling systems that are being used for the same purpose, and it is integrated in the sense that it forecasts both energy demand and gross domestic product simultaneously, taking into account the two way linkage existing between them. The energy economic system is formulated in a linear dynamic mode and Kalman filter is used for estimating the model. It is also applied for forecasting the state of the system in the future. The implementation of the model to the Greek energy economic system is presented as a case problem.     

A quadratic interval logit model for forecasting bankruptcy

This paper proposes a quadratic interval logit model (or quadratic interval logistic regression analysis) based on a quadratic programming approach to deal with binary response variables. This model combines the advantages of logit (or logistic regression) and Tanaka's quadratic interval regression model. As a demonstration, we applied this model to forecasting corporate distress in the UK. The results show that this model can support the logit model to discriminate between groups, and it provides more information to researchers.     

Futures literature and futures forecasting—A critical review

This paper presents a personal view of futures forecasting. Beginning with an overall view of the literature in the field with particular attention to the more significant contributions, methodology is then considered. Various general techniques of very long term forecasting are reviewed, followed by consideration of the particular fields of economic, social, technological, energy and politics. The social factor is seen to be all-pervading. A synthesis of futures forecasts is then attempted with particular reference to the British case. Finally, the significance of such forecasting approaches for the OR scientist is considered.     

A case study in econometric forecasting for alcoholic drinks

This paper reports on the estimation by econometric methods of demand equations for beer, spirits and wine, taken separately, using quarterly data over the period 1963 to 1978. The estimated demand equations are used to forecast demand over the 1979(1) to 1980(3) period for which actual data has recently become available, thereby providing a test, albeit not a very extensive one, of the short-term forecasting accuracy of the fitted equations conditional upon the values of the exogenous variables being known with certainty in the forecast period. The equations are also tested for stability. Finally, the paper considers the longer-term growth prospects for beer, spirits and wine consumption in the UK as the economy moves out of the current recession.     

The futility of forecasting

Long-range forecasting is an integral part of planning, but relying on its accuracy may be a mistake. The landscape is strewn with often wildly inaccurate forecasts. This article studies performances of some forecasts, analyses factors contributing to forecast error, and suggests ways in which management may deal with the uncertainty resulting from faulty forecasting performances.     

A comparison of encoding techniques

A limiting constraint of many management science techniques is that inputs from the decision maker based upon his experiences, opinions and intuition are not considered. For those models that do allow this type of input, it is assumed that they can be accurately and precisely defined in a subjective probability distribution. Little attention, however, has been directed towards evaluating the techniques to define these distributions in a management setting. This study investigates the relative merits of four of the most commonly used techniques for the quantification of subjective assessments. When these techniques were used with professionals whose jobs entail evaluation of uncertainty, a clear preference was shown. Additionally, some concluding observations concerning the selection and the application of assessment techniques are presented.     

A short term forecasting model for R & D expenditures

Whenever budget requests for individual R & D projects contain a built-in contingency then this will result in inefficient use of scarce resources. At the end of the year the annual budget may have been fully allocated but not necessarily fully utilized. This can result in delays or even complete cancellation of other proposed R & D projects because of an apparent lack of funding.The objective of the study was to forecast monthly R & D expenditures with a better accuracy than available from the subjective estimates of individual project leaders. The method described in this paper proved to be successful in practice.     

A short-term trend prediction model of topic over Sina Weibo dataset

Microblog has become a popular social network service. It provides a new communication platform for information acquisition, sharing and spreading. In addition to presenting daily-life reports from users, microblog also reports unexpected events, which get broad attention. How to forecast such unexpected events as early as possible? In this paper, we propose a short-term trend prediction model of topics in Sina Weibo, the most popular microblog service in China. Based on real microblog data, we first analyze which Weibo data attributes have influence on the spreading of topics, and then build a topic spreading model. Further, we develop a model of short-term trend prediction of topics. With dataset from Weibo, we test our algorithm and analyze the experimental data which shows that the proposed model can give a short-term trend prediction of Weibo topic.     

A critique of some techniques of manpower forecasting at national and sectoral level

This paper examines some of the principal methods which are currently being used by manpower planners in the U.K. and elsewhere to make projections of the demand for national and sectoral manpower in the medium term.Most of the paper deals with methods related to the Cobb-Douglas production function, and it is suggested that the “MRP” method developed by the OECD is, despite its apparent naïveté, the technique that offers the most promising basis for future development, in view of its theoritical soundness, its flexibility and its lack of restrictive assumptions.     

A hybrid ARIMA and support vector machines model in stock price forecasting

Traditionally, the autoregressive integrated moving average (ARIMA) model has been one of the most widely used linear models in time series forecasting. However, the ARIMA model cannot easily capture the nonlinear patterns. Support vector machines (SVMs), a novel neural network technique, have been successfully applied in solving nonlinear regression estimation problems. Therefore, this investigation proposes a hybrid methodology that exploits the unique strength of the ARIMA model and the SVMs model in forecasting stock prices problems. Real data sets of stock prices were used to examine the forecasting accuracy of the proposed model. The results of computational tests are very promising.     

Technological forecasting for power generation—A study using the Delphi technique

Technological forecasting is a powerful technique for obtaining insights into possible futuristic innovations. The forecast can provide a basis for long-range planning and policy formulation in organizations. Technological forecasting, though a useful and powerful tool, has yet to be applied widely in developing countries. This paper outlines a forecasting exercise carried out by the authors using the Delphi technique in India. The focus of the exercise was on electric energy generation and transportation. The findings have been compared with other available studies in India in similar areas. The paper also illustrates how additional information inputs can mould such forecasts. For this purpose, the authors have compared the results of the current study with the results of a similar study carried out in the same organization 8 years earlier. The change in perception regarding the timing of certain items is very evident and the underlying reasons have been given.