Week 9¶
Density Estimation(异常检测)¶
Problem Motivation¶
- Density Estimation Algotithm
- Anomaly detection example
Gaussian Distribution(高斯分布或正态分布)¶
The formula for the Gaussian density is:
$$
p(x) = \frac{1}{\sqrt{2\pi}\sigma}\exp\left(-\frac{(x-\mu)2}{2\sigma2}\right)
$$
- Gaussian distribution example
- Parameter estimation
Algorithm¶
- Anomaly detection example
Building an Anomaly Detection System¶
Developing and Evaluating an Anomaly Detection System¶
- the important of real-number evaluation
- Aircraft engines motivating example
- Alogorithm evaluation
Anomaly Detection vs. Supervised Learning¶
Choosing What Features to Use¶
- Non-gaussian features
- Error analysis for anomaly detection
- Monitoring computers in a data center
Multivariate Gaussian Distribution(Optional)¶
Multivariate Gaussian Distribution¶
- Motivating example:Monitoring machines in a data center
- Multivariate Gaussian(Normal)distribution
- Multivariate Gaussian(Normal) examples
Anomaly Detection using the Multivariate Gaussian Distribution¶
- Multivariate Gaussian (Normal) distribution
- Anomaly detection with the multivariate Gaussian
- Relationship to original model
- Original model vs. Multivariate Gaussian
Predicting Movie Ratings¶
Problem Formulation¶
- Example : Predicting movie ratings
Content Based Recommendations¶
- Content-based recommender systems
- Problem formulation
- Optimization objective
- Optimization algorithm
Collaborative Filtering¶
Collaborative Filtering¶
- Problem motivation
- Optimization algorithm
- Collaborative filtering
Collaborative Filtering Algorithm¶
- Collaborative filtering optimization objective
- Collaborative filtering algorithm
Low Rank Matrix Factorization(低秩矩阵分解)¶
Vectorization: Low Rank Matrix Factorization¶
- Collaborative filtering
- Finding related movies
Implementational Detail: Mean Normalization¶
- Users who have not reated any movies
- Mean Normalizatio
