MODELING AND MODELS
In business analytic, the term model is an abstraction of a real problem. In real life, when people try to solve a specific problem, they will always need to make a model to carry key features about that problem. The reason why they do this is because when a model can properly describe a problem, people can change the inputs to that model, and see whether the new inputs can solve the problem within the model. If it is pretty abstract, we can discuss it later in posts. But at least right now we know that making a model for a problem or a question is very crucial.
Now, since making model is important, there are several ways for you to make you own model.
1. Graphical Models
Graphical models are probably the most intuitive and least quantitive type of model.
This diagram indicates very clearly that how each feature or inputs affect each other in the models. It is quite easy to understand but it is hard to have quantitative details to help people while solving the problem.
2. Algebraic Models
Algebraic models are more specific to a problem because it is math formula which runs by numbers. Numbers represent quantitative details so basically, it can describe very specific relationship for each feature or input inside models.
Xj is the amount of product j produced
Uj is an upper limit on the amount of product j that can be produced
Pj is the unit profit margin for product j
aij is the amount of resource i consumed by each unit of product j
Bi is the amount of resource i available
n is the number of products
m is the number of scarce resources
Algebraic models are concise, but most of the time they are only suitable for people who know math pretty well.
3. Spreadsheet Models
Besides Algebraic models, excel can also do modeling. One of the benefits that using excel to do modeling is that you can check your model instantly. Since if you enter formula incorrectly in the model, you will immediately see an error. Then you can go back and fix.
However, the problem is you need to make a lot of practices on making spreadsheet models because Spreadsheet models require very accurate information and cannot have errors during the modeling process.
Thus, choose any models you like as long as you feel comfortable with.
Now after you choose a type of model that you want, then you can start the modeling process.
In general, there is a seven-step modeling process.
1.Define the problem
Every time, before doing any analysis you have to have a problem which leads you to the problem area to gather related data.
2.Collect and summarize data
Once you identify a problem, collect any data that are related to the problem. After you finish collecting data, you have to organize it into a format, filter those unrelated or useless data and summarize it. Now you have gathered all useful data for the problem and ready to proceed next step.
3.Develop a model
A model can be developed in a graphical way, an algebraic way or even in an excel sheet. No matter what kind of way you choose, the key point is to use your model to capture the important elements(input) of the business problem.
4.Verify the model
After you developed a model, we have to test the model so that we can know what is the accuracy level for the model. In other words, we need to determine whether the model developed in the previous step is an accurate representation of reality.
There are two ways to test:
- The analyst can use company’s current inputs as input parameters and calculates outputs. If the model’s outputs are the same or similar to the outputs that are observed by the company, the analyst can show the model can duplicate the current situation.
- Enter a number of input parameters and see whether the outputs are reasonable. One common approach is to input extreme values and see whether the output behave as they should.
If outputs are not correct, there could be two causes:
- The model is a poor representation of reality
- Our intuitions make us thinking the model outputs are wrong
5.Select one or more suitable decisions
If the model is correct, then we can make decisions and enter our decided numbers into the input parameters. We enter data several times, and we choose the best result.
6.Present the results to the organization
Usually, many people from management level do not really appreciate the model. However, you can include some of their people into the process of analyzing which make your model more believable.
7.Implement the model and update it over time
Finally, implement the model to reality.