In our last Part I blog, we introduced the capability of estimating key-input drivers based on early-stage metrics that represent requirements.  We even produced a CER based on a data-driven multiple regression, i.e., a linear model predicting Functional Complexity based only on two Key Performance Parameters (KPPs).

So how could you do the same?  Import the spreadsheet (given last time) into TrueFindings.  Note that spreadsheet row names now become the knowledgebase “column” fields.  Now per below, the first
tab-function “Distribution Finder” shows descriptive statistics for all or selected criteria—

The second tab-function “Dependency Finder” allow us to observe possible simple linear relationships
and the corresponding bivariate (Pearson) correlation coefficients—

As discussed in our Part-I blog, the third “Curve Finder” tab-function considers simple linear (or nonlinear) bivariate regressions.  Here we see the best one-predictor regression fits with R^2s, from the pool of requirements-metrics, e.g., KPPs—

Finally, specific to our objective, the fourth “MultiFinder” tab-function shows multiple regressions—

Note per the arrow above that a linear model now predicts Functional Complexity based on KPPs.  Specifically, it suggests that this key True-S input driver is best predicted by KPP#1 and KPP#2.  In fact, any additional KPP predictors are not statistically-significant, per inspection of the ANOVA p-values.

Spoiler Alert:  We could have gone directly to the MultiCurveFinder mode to evaluate best multiple-regression fits.  I wanted to show off the other tab-functions as well.  Also, keep in mind for this example, we deliberately chose pre-calibrated Functional Complexity values as the response (“y”) versus KPP requirements metrics (“x”s).  In fact, we could attempt to find relationships and equations between any of the field-variables.

What’s important to realize is the predicted values from an equation “Finding” are now immediately available within TruePlanning.  By deriving equations (hopefully based on calibrated data), you are essentially creating your own custom-calculator, as a pre-processor to our knowledge base algorithms!

This repeatable process allows us to perform Predictive Analytics to develop a method (i.e., a data-driven CER) that fine tunes TruePlanning for Software input driver(s) as a function of requirements, in this example KPPs.  {Of course, the same process would apply for analyzing Hardware inputs.}  In either case, you can leverage a calibrated knowledgebase based on your data to best fine-tune TruePlanning reflective of your past and future requirement metrics.

Success can be accelerated by the PRICE® Predictive Cost Analytics (PCA) integration of business, engineering, and program management objectives in order to produce competitive solutions.

TruePlanning immediately facilitates objective quantitative assessment of customer needs/
requirements, winning-business solutions, and project lifecycle cost management.

Contact me for a demonstration, either in person or via telecon.

We’ve had many success stories here.  You can be next!