Customer Value Optimization
Help service providers understand their customers to serve them better, thus Analytics is a step ahead of business intelligence (BI).
Recent advances in data collection techniques and database technologies have enabled organizations to store vast amounts of detailed information pertaining to their day-to-day business operations. Association learning is the process of discovering and processing interesting relations from large datasets. Very often the business finds value from rules derived from information that is beyond the simple view of transactions as groups of items (as derived from analysis of reports and dashboards).
Basic rules can be used to generate interesting knowledge between items from transactions. Nevertheless, it is imperative to extend the data using external knowledge about the items such as feedback, utilization, etc. to generate the rules. Association is used in understanding retail market basket analysis, where the aim is to discover which items are frequently bought together. They have the advantage that even if no prior relationships are given, rules can be generated that are highly unexpected and which would never have been specifically searched for, because they are inherently surprising. In addition analyzing the rules for only frequently occurring patterns can also be useful for anomaly detection, because those samples violating rules that usually hold are easy to identify and may be examples of interesting behavior. Association mining usually involves categorical data, but can be extended to apply for numerical data.
As the amount of data to mine for rules is usually large, efficient algorithms for generating association rules are necessary. This enables the user to maximize the quality of extracted associations and at the same time minimize the overall computation time required to generate the results.
Methods: Apriori, Bayes Networks, Markov Chains, Graphical Modeling.
- Sentiment analysis of customer feedback
- Customer experience and spare part consumption analysis
- Understanding the spread of epidemic diseases
- Network log analysis for better IT infrastructure security and availability