This blog uses the data mining client add-in for SQL Server to perform estimation analysis on Excel .
Article directory
You can first browse specific information about estimation analysis in the table of contents:
1. Estimation model
Estimation models extract patterns from data and use those patterns to predict factors that will affect outcomes. Results must be expressed in numerical values such as currency, sales, date or time.
For example, if the target column contains the school's graduation rate (expressed as a percentage), you can analyze factors that might increase or decrease the graduation rate, such as the number of students per school, the student-to-faculty ratio, and the number of teachers.
The Estimation Wizard uses the Microsoft Decision Tree algorithm. You can browse dependencies and patterns in an interactive viewer and quickly create graphs that represent more detailed information about the patterns discovered.
Data used in this analysis: 300 movie information (please leave a message in the comment area if you need resources)
2. Decision tree
Here I choose to use a decision tree for estimation:
general regressors are of numerical type, so some parameters cannot be included in the regressors:
The dependency network is as follows
According to the following animation, it can be seen that the least important parameter is POINTS
, and the most important parameter is BO
.
3. Neural Network
Another algorithm way:
After creating the structure, choose to add the model to the structure:
Then use the NBA data to act on the neural network algorithm:
Estimate the effect of the preceding parameters on the score per game:
4. Logistic regression
Choose a logistic regression algorithm: