_____Market aheadTM
The
MarketAhead program provides predictions for important
market variables, including the product demand and estimated profit.
The following variables are taken into account:
Sales
Unitary
price
Unitary cost
Market size
Publicity expenses
Consumer
income
Seasonal index
Profit
Market elasticities with
respect to price, publicity and consumer income
Fixed costs
Two
kinds of preditions are calculated:
Predictions based on
historic data where the analysis of time series is implemented.
The
input to the program is a sequence of values, supposed to be the
values of a variable of interest taken from a stochastic process.
These may be, for example, the demand for a product for consecutive
days, weeks, months etc. during some past time interval. MarketAhead
looks for tendencies in the series and for possible seasonal
(periodical) changes. Once an approximating curve is found, a
prediction for the future is produced. If the provided data is
charged with random errors or uncertainty, you can get the
predictions with corresponding confidence intervals for the
prediction value, as function of time. The data can be typed and
stored in a file, retrieved, edited or imported from an ASCII text
file.
Demand
and profit predictions based on a market model.
The
market model is of "exponential" type involving model
elasticity with respect to price, advertising and consumer income.
First, forecasts for all market variables except demand, sales and
profit are generated using time series analysis. Then, these
forecasts are used as input to the market model to generate the
future demand and profit. This makes the predictions consistent with
the market model, and prevents from contradictions and logical errors
in the projections. So, the demand projection can be obtained both as
the result of the time series analysis of the demand, and as the
result of market model application. The two projection should
coincide or be close enough to each other. If this is not the case,
the input data, as well the used model parameters should be verified.
Uncertainty analysis. Uncertainty for predictions of kind 1 and 2 is analyzed. The user is asked to assess the standard deviation SD of possible errors in the historic data. Then, the program repeatedly generates new data charged with random deviations due to the given SD, and produces the series of prediction trajectories. The trajectories are being stored and analyzed. Then, the corresponding confidence interval for the predictions are calculated, for a given confidence level. The confidence intervals are shown as functions of time.
MarketAhead runs on a PC with Windows XP or later.
Historic data and projection for a market variable with saturation. Sigmoid curve fit.
A prediction of a variable with periodical seasonal component.
Uncertainty analysis
The data obtained from a real market are always charged with some uncertainty. What the program does is to generate multiple data sets where the original data are changed randomly, with a given standard deviation and to calculate the corresponding projections. This is a simple procedure without the use of any sophisticated probabilistic approach. However, the algorithm is robust and may provide a useful information on the future market behavior and its sensitivity to the input data. A set of 200 possible projections is generated. It is assumed that the probability distribution for the projecton for any fixed time instant is normal. The user declares a probability level P, and the program calculates the possible confidence intervals. By the confidence intervals we mean the intervals in which the expected projection belongs with probability P. The results are shown in graphically.
Calculating uncertainty
Uncertainty : confidence intervals
Plot flow: the plots move, can be selected with a click
Download a free demo here
BLUESSS simulation package : queuing, continuous simulation
Slicer : Converts STL files into slices
http://www.youtube.com/watch?v=4me_lB7EEW0&feature=player_detailpage Elevator simulation