I would like to cluster user into at least 3, and at most 7 segments based on different credit and paying back variables that i would collected from two different databases. In the attached file, i put an example of what table I will have to do this. The user_id (A-E) represents my users which, in reality, is about 20, 000. I put different potential variable that would be part of clustering. I would assume they can rage from 5-15 variables (in the example, dunning state, no. of missing date, and so on).
My question is,
1) Is it possible to do a k-means clustering to achieve such segmentation? Or do I need to have other statistical means I should purse?
2) Which software can do the job well (R or Spss)?
Excuse any nativity on my part, as i am new to quantitative data analysis.
Very interested question you have raised,
There is solution to solve your doubt and kindly try to follow the procedure:
1. First reconstruct the missing data in the dataset using Dynamic Data Cleaning Technique (Refer the Paper: " Article A New Dynamic Data Cleaning Technique for Improving Incomple...
"), i hope this concept is very helpful to reconstruct the missing data points in the dataset.2. Use the DAAC (Dynamic Automatic Agglomerative Clustering) scheme to automatically split the pre-processed dataset into finest number dissimilar clusters without user input (Refer Paper: " An Improved Frequency Based Agglomerative Clustering Algorithm for Detecting Distinct Clusters on Two Dimensional Dataset ")
3. Can easily evaluate the clustering result of user dataset using cluster validation scheme (Refer Paper :An Improved Frequency Based Agglomerative Clustering Algorithm for Detecting Distinct Clusters on Two Dimensional Dataset)
Herewith, i hope you can follow the above given procedure and identify the distinct groups over the user dataset for you analysis.
Thank You
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps