Getting started with Time Series Forecasting
If we are solving a supervised learning problem, there we have independent and dependent variables so what we can do is apply a function on how variables are correlated with one another and then we can go for the analytics and model-building part, but in time-series analytics you just have one variable that is time.
For example, you are a car showroom, which is very famous in your city, so what do you try to see is how many cars you sell every month or a year, in order to do that what you will do is add the number of cars you sold each month and then sum it so you will have all the data till the present month but what if you want to know the sales the next month or the next year. Now think, you just have one variable that is the number of cars sold and you need to predict that variable over time, for doing so you will need time series forecasting.
What is time series?
A time series is a series of data points indexed in time order. Most commonly, a time series is a sequence taken at successive equally spaced points in time. Thus it is a sequence of discrete-time data. -Wikipedia
Some Applications of Time Series
- Sales forecasting
- Stock market forecasting
- User behaviour analysis
- Weather forecasting
- Census analysis
- Seismological predictions
Components of Time Series
- Trend— when the time series analysis shows a general pattern that is upwards we call it an uptrend also if the trend exhibits a lower pattern that is down we’d call it a downtrend and if there was no trend we call it a stationary trend.
- Seasonality — when the time series shows upward or downward swings which are repeating over a fixed period of time. so it’s a repeating pattern within a fixed time period.
- Irregularity — when the time series shows sudden changes which are unlikely to be repeated which are also known as residual or random components.
- Cyclic —when time series keeps repeating up and down movements so over more than a year so they don’t have a fixed pattern so they can happen anytime let’s say in two years then the fourth year than maybe in six months so they keep on repeating and they are much harder to predict.
