Time Series discrete note

Anh-Thi Dinh
draft
⚠️
This is a quick & dirty draft, for me only!
Terminologies & fields of research
  • Burst detection: An unexpectedly large number of events occurring within some certain temporal or spatial region is called a burst, suggesting unusual behaviors or activities.
  • Time Series Regression: (ref) Time series regression is a statistical method for predicting a future response based on the response history (known as autoregressive dynamics) and the transfer of dynamics from relevant predictors. Time series regression can help you understand and predict the behavior of dynamic systems from experimental or observational data. Time series regression is commonly used for modeling and forecasting of economic, financial, and biological systems.
  • Time Series Classification: (ref) Time series classification deals with classifying the data points over the time based on its' behavior. There can be data sets which behave in an abnormal manner when comparing with other data sets. Identifying unusual and anomalous time series is becoming increasingly common for organizations
    • (ref) Time series classification data differs from a regular classification problem since the attributes have an ordered sequence.
  • Anomaly Detection:
    • A part in the same time series.
    • Finding one or more time series which are different from others.
    • Some abnormal points in the same time series.
    • Applied for both univariate and multivariate time series.
Read_CSV
More here.
1df_13 = pd.read_csv(path_file,
2                    index_col='timestamp',
3                    parse_dates=True, # index contains dates
4                    infer_datetime_format=True, # auto regconize format
5                    cache_dates=True) # faster
Find the windows of time series
Suppose we have data like in below, we wanna find the common length interval of all groups.
There are other cases need to be considered,
The gaps are not regular
If we choose the gaps (to determine the windows) too small, there are some windows have only 1 point like in this case.