Markov knows the future
Forget fortune tellers and fortune cookies, only Markov can help you make reliable forecasts. It supports you making better decisions. The technique is called after the Russian mathematician Andrei Andreyevich Markov, a genius living around the nineteen hundreds. Although Markov is still relatively unknown the technique is an immense powerful one. It has many applications in the business world. Markov is useful in any application were you would like to make a prediction. For example:
- The google search engine uses Markov to help decide what search result you are most likely most interested in. That is how they rank the results.
- Speech recognition software uses it to help understand which words were most likely spoken.
- HR departments use Markov to predict the flow of people in their organisations according to their job positions. What is the probability that a junior engineer in the organisation one day becomes a manager, or even the CEO. This helps HR departments deciding whether to focus on finding people internally for job position or to go and look outside the organisation.
- The stock market uses it to predict the closing prices of stocks. No, Markov is not to blame for financial crisis, a bad model may be yes.
- Competitive Intelligent department use it to predict market shares based on current market distribution between brands.
- The functional safety industry uses Markov to predict the probability of fail dangerous (PFD) of safety systems.
- The banking industry uses it as a Risk Management tool. Can this customer pay back the loan amount he or she is applying for? What is our risk exposure today with the outstanding loans we have?
- The weather department uses it to forecast tomorrows weather.
The number of applications are endless. All you need is a tool that helps you create a model and the rest is history…
Introducing the Markov Master Tool
Interested in the Markov Master Tool?
Contact us and we will contact you
- Timeless engine. The transition rates can be in milli seconds, seconds, minutes, hours, days, months, etc
- Unlimited number of states and transitions
- Unlimited groups combining states of interest
- Unlimited data sets to make multiple calculations with the same model but different data
- Unlimited redistributions allowing you to move probabilities from one state to another state at any time
- Tabular and graphical model creation
- Graphical results and numerical results
- State and group analysers helping you understand why the results are what they are
- Support website with manual and forum
- Option to export the results in cvs or report in PDF format
- Runs on Mac and Microsoft
The states and transitions you need to create for your model can be entered in two ways. First of all via a graphical model. The graphical model works well for small models. It allows you to create bubbles representing the states and to drag transitions between the states. You can give each state a colour showing clearly which states belong together.
When the models are getting bigger the graphical solutions is not the preferred method to create the model. For those medium to large size models we recommend the tabular format. You will be surprised how fast you can create models this way.
If you still prefer the graphical way no worries. The tabular format and graphical format are link with each other.
Once the model is ready it usually does not change that much any more. What does change is the data you want to use to make calculations.
The tool allows you to create unlimited data sets per model. Each data set has the same variables but you can change the values of the variables as you wish. This feature is very useful if you, for example, use different standards with data, or if you have data that belongs to different time periods, e.g., from 2012, 2013, 2014, etc . Now you can make very easily very quick comparisons.
When you want to perform the calculation you should select the data set you want to use and the tool calculates the results. Keep adding calculations using different datasets and you have will have enough information to make your decisions effectively.
Markov calculates the probability of being in a State at any given time. When you evaluate your model over time the probabilities in the different states keep changing slowly. Sometimes you need to model a “big” change though. Lets take as an example the reliability of a machine. Over time the probability that the machine will be in a failed state will increase. But what if you do every year very thorough maintenance. How does the probability change?
For those models where you need to change the probability of being in a state at any given time there is the redistribution function. You decide how much probability you take out of which states, and you decide how you distribute them over any of the other states. Very powerful.
Results and Reporting
You can view your calculation results in different ways. The tool allows you to see the results in a graphical way or as raw numbers. The results can be exported into cvs format so you can use the raw data in your favorite spreadsheet program or with any other application you want use. You can also view online and print a report in PDF format.