From the course: Machine Learning in Telecommunication: From Basics to Real-World Cases

Unlock this course with a free trial

Join today to access over 24,800 courses taught by industry experts.

Anomaly detection: Spotting outliers in network performance

Anomaly detection: Spotting outliers in network performance

From the course: Machine Learning in Telecommunication: From Basics to Real-World Cases

Start my 1-month free trial Buy for my team

Anomaly detection: Spotting outliers in network performance

“

- [Instructor] To identify the outliers in the dataset, we sometimes use Z-score. What if your phone bill suddenly skyrocketed one month? That would be weird, right? That's exactly what anomaly detection is all about. And we use the term Z-score to measure how far the point is from the average value. So it is similar to what we talked about earlier. We talked about standard deviation and we decide that anything which is any data point, which is far away from the mean value, would be determined by Z-score. If it is 2.5, it means that it is 2.5 standard deviations away from the mean, which is signifying that it is not closer to the mean, it is bit far off. And even if it is three or more, it means that it is an outlier. So it helps us to identify the outliers or detect the anomalies in the data. And there are certain examples where we can use it. For example, for the data set where we have different values, you would like to see what is the outlier here? You can easily see it is 100…

Contents