Forecast Module

Qualitative forecast is the key to solving of the most actual business challenges, as optimization of the amount of stocks in storage or financial flows, budgeting, assessment of investment attractiveness and more.

In practice it is very difficult to get qualitative forecast making just one simple operation. Mostly few steps should be done to solve this problem as:

1. Connection of data sets
2. Cleaning data
3. Data Transforming
4. Model building
5. Prediction

BMS Soft Forecast Module includes the full set of tools for solving this problem, starting from connection to different data sources and ending with the construction of adaptive models. The principles are implemented in this module and provide support for the forecast process at all stages of process.

Typically the hypotheses suggestion concerning the effects of factors on the end result is made by the expert, who mostly relies on his experience and intuition, though our Forecast Module includes tools applied for preliminary analysis: 

• • Primary hypothesis control
• • Assessment of the auto correlation factors importance
• • OLAR Cubes
• • Charts and histograms

Forecast is one of the most essential and, most difficult tasks. Its difficultness is determined by many reasons:

• • Lack of data quality and quantity
• • Inconstancy (uncertainty)of environment, where the process runs
• • The impact of subjective factors

While forecasting it is necessary to solve out the regression tasks, i.e. to forecast a continuous output field value based on several input parameters. That’s why there are such models types in the BMS Soft Forecast Module as:

1. Regression analysis
2. Neural network
3. Expert system
4. Models built on the basis of "nearest neighbor" method

  After the forecast model creation the forecast itself can be got easily based on data entered manually or downloaded from external sources.

Forecast module interface

The example of examination of logistic regression.

model results display