What would happen if we perform Time Series Decomposition on a data that only has trend and irregular components? I am performing Decomposition of a Timeseries data. It appears that the data does not have a seasonal component and possess only trend and irregular components. However on still performing the decomposition analysis, seasonal component throws me values between 0.88 and 1.11 (which is close to 1). Please note I am performing a Multiplicative Decomposition analysis. Can some only tell me if this range of seasonality values indicates if a seasonality is present in the data?
 A: A little too statistic and subjective answer! What I find is that it is not the exact value of the seasonal component, but the below 3 key factors that would help whether you would have a seasonal pattern.
Points to consider from an exploratory perspective


*

*The visual pattern you observe in the data 

*The business behind the data (e.g: Sales data which may definitely have a pattern over time)

*The length of the data to find a repetitive pattern


Points to consider from an statistical perspective


*

*Decomposition is not the only way to find seasonality. There are numerous time series methods to analyze a ts data

*Consider using a double exponential smoothening (Holt's smoothening) using R if your sure that you do not have seasonality. If you suspect seasonality, try the Holt-Winter's smoothening which has 3 smoothening parameters (1 for level, 1 for trend and 1 for seasonal). If the code error's out, then it means you do not have one of the parameters in your dataset. mostly this would be the seasonal component that would be missing and level and trend are seen very commonly in most ts data. 

*Through ARIMA, one can also see the ACF and PACF plots for occasional spike that would indicate seasonality. 


Any other thoughts or suggestions on this?
