Time is the most important factor which ensures success in a business. It’s very difficult to keep up with the pace of time.  But, technology has developed some powerful algorithms using which we can see things ahead of time. Don’t worry, I am not talking about any Time Machine. Let’s be realistic here!

I’m talking about the algorithms of prediction & forecasting. One such method, which deals with time based data is Time Series Algorithm. As the name suggests, it involves working on time (minutes, hours, days, years) based data, to derive hidden insights to make informed decision making.

What is Time Series Forecasting?

Time-series forecasting is a quantitative forecasting   technique. It measures data gathered over time to identify trends. The data may be taken over any interval   yearly, monthly, daily, hourly or longer. Trend, cyclical, seasonal and irregular components make up the time series.

The trend component refers to the data’s gradual shifting over time. It is often shown as an upward and downward sloping line to represent increasing or decreasing trends.

Cyclical components lie above or below of the trend line and repeat for a year or longer period of time. The business cycle represents a cyclical component. Seasonal components are similar to cyclical in their repetitive nature, but they occur in one-year periods. And Irregular components happen randomly and cannot be predicted.

The process of Time Series Algorithm

The Time Series Algorithm is generally used when historical data plays an important role in predictive future forecasting. The process is shown in this diagram.

Let me briefly explain these steps

Determine Trend

Trend is increase or decrease over a period of time. This can be caused due to upward or downward pattern. Example which cause this are population increase, technology adaption etc.

Graph shown here is a time series without trend and time series with trend

Determine Seasonality

Seasonality means that a fluctuation occurs over certain periods of time in each year or long period of time.

here is a graph of time series which shows without Seasonality and time series with Seasonality. As you can see that in time series with Seasonality, there is increase in value observed in months 7 and 8

Box plots are very useful to visualize seasonality

Determine Auto-correlation

Auto-correlation is nothing but how a point in time-series impacts future values in time-series algorithm.

This can be determined using a Partial Auto-correlation(PACF) plot

Time-Series Modeling

Time-series modeling tries to fit a model to old values, so that it can be used to forecast future values. One of the most commonly used algorithms for time series modeling is ARIMA

This algorithm first converts the data in to information, what we call it a stationary time-series, by removing the trend and seasonality. Then it tries to fit a model on the accurate data

The analysis done about trend, seasonality and auto-correlation earlier in the process serve as input parameters for ARIMA Forecasting

Once the model is ready, it can be used to Forecasting the future values

Large-Scale Price Modeling and forecasting with Aster

• Many business situations require price-forecasting to be scalable. A retailer may have thousands or millions of products for which sales forecasting is required.
• A telecom company would typically have thousands of voice and roaming routes to forecast price for voice call.
• product or product group has its own time-series model, so it would be required to train various time-series models in parallel

Aster, with its scalable algorithms, has capability to make large-scale forecasting on various models in a continues way

Some of the Aster algorithms used are

1. Aster GLM to determine trend and seasonality
2. Aster-R or Aster Streaming for partial correlation
3. Aster ARIMA for time-series modeling
4. ARIMA Predictor for time-series forecasting

So hopefully this blog gives you an idea on how forecasting and time-series modeling Algorithm works, as well as how Aster capabilities used in forecasting. Please feel free to comment