Data Mining is a process which is used to retrieve a pattern that describes data from the data
Data mining uses well-researched statistical principles to discover patterns in your data. By applying the data mining algorithms in Analysis Services to your data, you can forecast trends, identify patterns, create rules and recommendations, analyze the sequence of events in complex data sets, and gain new insights.
- Tabular data sources like TextFile, spreadsheet and OLAP cube created in SSAS can be used for Data mining whereas in-memory databases can’t be used.
- We can find patterns in data by using different algorithms such as Neural networks, Decision Trees and Clustering and can also create own custom plug-in algorithms.
Follow the steps below to do Data Mining :
Step1: Create an SSAS Multidimensional and Data mining Project and build a cube.
Here I have created an SSAS project Cubeanalysis and also created Data Source, Data Source Views, and Cube
Step2: In Solution Explorer right Click on Mining Structures select New Mining Structure.
Step 3: A Data Mining Wizard will pop up select From the existing cube option it will pull the data from the cube.
Step 4:Create the Data Mining Structure it lets you select the mining technique to retrieve the patterns from data there are two different option One lets you select the inbuilt model other is to create it without models
We are going to select the first option and select Microsoft Clustering and Click next.
Step 5: Select one dimension from the dimension here I have selected Dim Employee
Step 6: Select a Case Key which means select a key value from the dimension selected
Step 7: Select the measures and attributes you want to include in the mining model.
Step 8:In Select the Column usage is used to select the columns that you want to use as input and predict In this example we will organize the information in clusters and all the data will be used as input.
Step 9:In Columns Content and Data Type if the columns are discrete then convert them into Continous cause it can throw an error if discrete
Step 9:Create Testing Set let the percentage of data testing be 30% and click next
Step10: Give an appropriate name for the Mining structure and Mining model Click Finish.
Step 11: In solution explorer right click on the mining structure that is created and click on the process it will build and deploy the mining model and then run the mining model.
Step 12:Once the process succeeded Click close
Step 13:Click on the Mining Model viewer to view the cluster diagram that represents the data here if you use more complex data this Cluster diagram will get more complex
Similarly, you can also view the Cluster profiles which shows values for every column.
Start with MSBI video as mentioned below to learn topic practically in-depth
