Agenda : • Time Series • Importance of Time Series • Smoothing Methods • Project on Time Series with Airlines dataset • ARIMA Model
Time Series :
- It is an sequence of values of a variable at equal spaced time internal .
- It is a series of datapoints ordered in time .
- Time series analysis is a statistical technique that deals with time series data ,or trend aalysis.Time series data means that data is in a series of particular periods or intervals.
- We refer today as T ,yesterday as T-1 day before yesterday as T-2 and henceforth .And T+1 for tomorrow and so on .
- The data is considered in 3 types:
- Time series data : A set of observation on the values that a variable take different times .
- Cross-Sectional data: Data of one or more variables, collected at the same point in tie
- Pooled data: A combination of time series data and cross-sectional data .
- We dont take in consideration seasonal data or stats in time series .
- We drop trends also in time series because it get vulnerabilities .
- After dropping these factors we can do time series forecasting .
- We can have different types of time interval .
- Here have only 2 variable time and value. Importance of Time Series :
- Time series is very important to solve a lot of problems in the business.
- Based on time we create a lot of data .
- We can use this to predict future operations. Application :
- Economic Forecasting
- Sales Forecasting
- Budget analysis
- Stock market analysis
- Inventory studies etc
Time series components:
- It can be described in terms of 4 basic classes of components:
- Trend : It is a long term direction of a time series .It exist in long term increase or decrease in the data. It does not have to be linear, sometimes we will refer to a trend “changing direction”, when it might go from an increasing trend to a decreasing trend.
- Seasonal : It is a regular pattern of variability within certain time periods such as year.
- Cyclical: Any regular pattern of sequences of values above and below the trend.
- Irregular: Cannot be defined it can change without a pattern.
Stationarity :
- Data should be stationary..
- By nature no data has stationary format, We have to convert the data into stationary data then we can apply TimeSeries techniques.
Smoothing Methods:
- It removes the random variations and shows trend and cyclic components.
- When a time series contains a large amount of noise, it can be difficult to visualize any underlying trend.
- There are 2 methods. ARIMA Model:
- Auto Regressive (AR) Integrated(I) Moving Average (MA) ACF &PACF :
- Auto Correlation Factor and Partial Auto Correlation Factor