Evaluation of aggregate and individual forecast method selection rules

Article Abstract:

Univariate forecasting methods are used in production control and are essential to modern manufacturing methods The choice of a forecasting method is critical to production control, particularly to manufacturers with large numbers of production processes and product lines. There are two forecasting approaches: aggregate selection and individual selection. Aggregate forecasting applies a single method to the entire series of production processes. Individual forecasting determines the appropriate method and applies it to each production process, and is an expensive, time consuming process characterized by complex computations. An analysis of 263 data series from a single organization reveals that individual forecasting is a more accurate method for forecasting production processes with long lead times and sample variability, but does not always give superior results due to the lack of stability in data series variables. Additionally, results reveal that aggregate forecasting is a cost-effective compromise that, when applied across a wide cross-section of series and times, is simpler and of comparable accuracy to individual forecasting.

Research, Forecasting, Comparative management

---

Diffusion and advertising: the German telephone campaign

Article Abstract:

The effect of advertising on new telephone diffusion in West Germany is examined. Alternate ways of integrating advertising with the 'Bass-diffusion' model are offered and tested empirically. Econometric investigation provides results consistent with a behavioral assumption: advertising mainly affects the demand of imitators. The lag structure entailed in advertising response is best measured by a 'Nerlove-Arrow' model. The maximum effect of advertising is reached several months after it is presented. Seventy-two percent of total advertising effect is shown to occur between four and nine months after advertisements appear.

Models, Telecommunications services industry, Advertising, Science and technology policy, Marketing, Marketing models, Diversification in industry, Industrial diversification, West Germany, Germany, West

---

The problem of dimensionality in stratified sampling

Article Abstract:

Stratified sampling is a natural technique for variance reduction but the high dimensionality of sample space often skews its results. The use of shadow response variables, those variables that have similar properties of the control variables, can be used to significantly reduce sample space variation to within a reasonable tolerance. The selection of good shadow response variables is necessary for successful stratified sampling.

Samples (Merchandising), Samples (Products)

Similar abstracts:

• Abstracts: An empirical evaluation of alternative forecasting combinations. A state space modeling approach for time series forecasting
• Abstracts: Expectation formation and financial ratio adjustment processes. Measurement of financial leverage in the presence of unfunded pension obligations
• Abstracts: Determinants of actuarial cost method changes for pension accounting and funding. Understanding the accounting for defined benefit pension plans
• Abstracts: A comparison of slack resources in high and low performing British companies. Excellent companies as social movements
• Abstracts: Organizational decline and innovation: a contingency framework. Temporal sequences in information acquisition for decision making: a focus on source and medium

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.