What is the relationships between data?
In our data, there are hidden links. For example, a digital marketer will find it evident that there is a relationship between the budget for a Google Ads campaign and the number of leads generated.
Other connections might not be so obvious. Is there a relationship between the weather and the number of sales in a bakery? Is there a relationship between the number of shops in a neighborhood and the selling price of houses?
These are all one-to-one relationships and seem apparent. But in a typical dataset, there are several parameters, and the relationship between a “goal” and the various parameters are not that straightforward.
What is a predictive Model?
TADA can connect the dots between these parameters. It understands the dependencies between:
- the selling price of a house,
- the number of shops in the neighborhood,
- the location,
- the year of sale.
Thanks to its powerful Machine Learning algorithm, it learns from existing data and finds correlations between them. It expresses what it has learned through an equation where the goal is the result, and the influencing criteria are the parameters. This equation is the predictive model.
Once it has created this equation, it can input new data into the equation. The result of the equation is the ‘goal,’ i.e., what we want to predict.
Does a Unique Predictive Model exist?
TADA has a way of learning which makes it try out several correlations between the data to create the equation, i.e., the predictive model. It explores different options. Each time it works on generating a new model, it explores other options.
Therefore, each time it runs, it generates a different model. Some of the models are very close to each other and some are very far apart. There is no Unique Predictive Model. There can be several different models for each dataset.
How do we know which parameters influence the most the predictive model, i.e. the prediction? Read our article about Sensitivity Analysis to find out.