What is the Goal of Data Mining?
A list of ten different "goals" — predicting sales, detecting fraud, personalizing recommendations, optimizing routes — makes data mining sound like ten different activities. It's really the same handful of underlying functions, covered in full in the lesson on categories of functions in data mining, aimed at different business objectives. This lesson makes that mapping explicit: a goal is the business "why," a function is the technical "what kind of pattern," and confusing the two is exactly why lists like this one tend to feel repetitive without actually being organized.
Mapping Goals to the Functions That Deliver Them
| Commonly Listed Goal | Underlying Function | Covered In |
|---|---|---|
| Predictive analysis (sales, loan risk) | Classification / Regression | Lesson on categories of functions; lesson on classification by backpropagation |
| Identifying patterns and relationships | Association / Correlation | Lesson on association rule mining; lesson on correlation analysis |
| Anomaly detection (fraud, cyberattacks) | Outlier detection | Lesson on data mining applications (worked z-score example) |
| Customer segmentation | Clustering | Clustering lessons throughout this series; lesson on data mining for business intelligence |
| Personalization and recommendations | Association / Collaborative filtering | Lesson on data mining projects; lesson on data mining applications |
| Scientific discovery | Clustering / Sequence mining | Lesson on trends in data mining |
| Process optimization, inventory, routing | Regression / Clustering applied to operations | Lesson on the strategic value of data mining (cost-reduction example) |
| Risk management (credit, insurance) | Classification | Lesson on data mining applications (credit scoring) |
Every row in this table reduces to the same small set of functions covered once, in depth, elsewhere in this series — the "goal" is simply which business problem that function gets pointed at.
Ethics Is a Constraint on Every Goal, Not a Separate Goal
It's worth correcting one more categorization issue: privacy protection and bias avoidance are frequently listed as if they were another goal alongside prediction and segmentation, but they aren't a goal in the same sense — no organization sets out to "do data mining for the purpose of privacy protection." These are constraints that apply while pursuing any of the actual goals above, covered in full in the lesson on the challenges of data mining, and treating them as a tenth item on a list of objectives understates how directly they shape whether a project pursuing any of the other nine goals can be carried out responsibly at all.
Why the Underlying Function Matters More Than the Stated Goal
Two organizations with completely different stated goals — a bank managing credit risk and a streaming service personalizing recommendations — can end up using the exact same underlying function (classification, in the bank's case predicting default risk, in the streaming service's case predicting which show a viewer will finish watching). This is precisely why this series is organized around functions and techniques rather than around goals: understanding how classification actually works, once, transfers immediately to any goal that happens to need it — which is a far more durable kind of knowledge than memorizing ten separately named "goals" that all reduce to the same handful of underlying mechanisms.