Difference between revisions of "...find outliers"
| Line 1: | Line 1: | ||
* [[AI Solver]] | * [[AI Solver]] | ||
| + | ** Looking for event based? | ||
| + | *** [[Bidirectional Long Short-Term Memory (BI-LSTM) with Attention Mechanism]] | ||
| + | ** Do you have > 100 features? | ||
| + | *** Yes, then try [[One-class Support Vector Machine (SVM)]] | ||
| + | *** No, need fast training, then try [[Principal Component Analysis (PCA)]]-based Anomaly Detection | ||
| + | ** Also consider... | ||
| + | *** [[K-Means]] Clustering | ||
| + | *** [[Hierarchical Clustering; Agglomerative (HAC) & Divisive (HDC)]] | ||
| + | *** [[One-class Support Vector Machine (SVM)]] | ||
| + | **** [[Variational Autoencoder (VAE)]] | ||
| + | *** [[Autoencoder (AE) / Encoder-Decoder]] | ||
| + | *** [[Radial Basis Function Network (RBFN)]] | ||
| + | |||
| + | * [http://en.wikipedia.org/wiki/Outlier Outlier | Wikipedia] | ||
* [http://pdfs.semanticscholar.org/4c68/4a9ba057fb7e61733ff554fe2975a2c91096.pdf Outlier Detection and Removal Algorithm in K-Means and Hierarchical Clustering | A. Barai and L. Dey] | * [http://pdfs.semanticscholar.org/4c68/4a9ba057fb7e61733ff554fe2975a2c91096.pdf Outlier Detection and Removal Algorithm in K-Means and Hierarchical Clustering | A. Barai and L. Dey] | ||
* [[Capabilities]] | * [[Capabilities]] | ||
| Line 6: | Line 20: | ||
** [[Signals]] | ** [[Signals]] | ||
** [[Pathology]] | ** [[Pathology]] | ||
| − | |||
Anomaly detection. Sometimes the goal is to identify data points that are simply unusual. In fraud detection, for example, any highly unusual credit card spending patterns are suspect. The possible variations are so numerous and the training examples so few, that it's not feasible to learn what fraudulent activity looks like. The approach that anomaly detection takes is to simply learn what normal activity looks like (using a history non-fraudulent transactions) and identify anything that is significantly different. | Anomaly detection. Sometimes the goal is to identify data points that are simply unusual. In fraud detection, for example, any highly unusual credit card spending patterns are suspect. The possible variations are so numerous and the training examples so few, that it's not feasible to learn what fraudulent activity looks like. The approach that anomaly detection takes is to simply learn what normal activity looks like (using a history non-fraudulent transactions) and identify anything that is significantly different. | ||
| − | |||
| − | |||
| − | |||
| − | + | <youtube>086OcT-5DYI</youtube> | |
| − | + | <youtube>MUcrGtLKK7I</youtube> | |
| − | + | <youtube>QaVL4Ht3u8w</youtube> | |
| − | + | <youtube>LRqX5uO5StA</youtube> | |
| − | |||
Revision as of 10:23, 7 January 2019
- AI Solver
- Looking for event based?
- Do you have > 100 features?
- Yes, then try One-class Support Vector Machine (SVM)
- No, need fast training, then try Principal Component Analysis (PCA)-based Anomaly Detection
- Also consider...
- Outlier | Wikipedia
- Outlier Detection and Removal Algorithm in K-Means and Hierarchical Clustering | A. Barai and L. Dey
- Capabilities
Anomaly detection. Sometimes the goal is to identify data points that are simply unusual. In fraud detection, for example, any highly unusual credit card spending patterns are suspect. The possible variations are so numerous and the training examples so few, that it's not feasible to learn what fraudulent activity looks like. The approach that anomaly detection takes is to simply learn what normal activity looks like (using a history non-fraudulent transactions) and identify anything that is significantly different.
