- Year not given · 2+6 marks
- Year not given · 3+5 marks
DM techniques; likelihood and distance principle
Short answer
Data mining techniques include classification, prediction or regression, clustering, association rule mining, sequential pattern mining, anomaly detection, summarisation and time-series analysis. Likelihood and distance principles mean nearby objects in attribute space are alike, so nearest neighbours or highest-likelihood groups decide classification or clustering. Example: customer X at (26,42) is nearest to two laptop buyers, so X is predicted yes and clustered with them.
What are the different techniques used in data mining? Explain likelihood and distance principle ("Data that are close to each other are very alike") with suitable supporting example.
Data mining techniques
- Classification: assign objects to predefined classes using a model learnt from labelled data (decision tree, Naive Bayes, KNN, neural network).
- Prediction/regression: estimate a continuous value.
- Clustering: group unlabelled objects into natural clusters (k-means, hierarchical, DBSCAN).
- Association rule mining: find items that occur together (Apriori, FP-growth), e.g. {bread} → {butter}.
- Sequential pattern mining: frequent ordered events over time.
- Anomaly (outlier) detection: find unusual objects, e.g. fraud.
- Summarisation/characterisation and discrimination: concise description of a data set and comparison of classes.
- Time-series and trend analysis.
Likelihood and distance principle
"Data that are close to each other are very alike": the distance between objects in attribute space tells how similar they are, and objects near to each other probably have the same class or belong to the same cluster. The likelihood part chooses the most probable group for a new object, given its neighbours.
Example: customers described by (age, income in thousand rupees).
| Customer | Age | Income | Buys laptop |
|---|---|---|---|
| C1 | 25 | 40 | yes |
| C2 | 27 | 45 | yes |
| C3 | 55 | 90 | no |
New customer X = (26, 42). d(X, C1) = √(1 + 4) = 2.24, d(X, C2) = √(1 + 9) = 3.16, d(X, C3) = √(841 + 2304) = 56.1. The two nearest neighbours buy laptops, so X is predicted "yes". The same principle puts C1, C2 and X in one cluster and C3 in another.
✓ Answer checked by Claude Opus 5.5 (high effort)
Related questions
- Short note: Likelihood and distance · Data Mining, chapter 3
- Likelihood and distance; K-means drawbacks · Data Mining, chapter 3
- Web mining, importance and techniques · Data Mining, chapter 7
- Data mining and its applications · Data Mining, chapter 1
- Data mining vs DW and applications · Data Mining, chapter 1
- CRM; security and privacy of data mining · Data Mining, chapter 7
- Data mining and business knowledge · Information System, chapter 4
- Data warehouse and data mining · Information System, chapter 4 · asked 2×
- Short note: data mining and warehousing · Information System, chapter 4
- Data transfer techniques for slow I/O · Instrumentation II, chapter 1 · asked 3×