Exponential smoothing:

This method is sophisticated weighted moving average method that calculates the average of a time series by giving recent demands more weight than earlier demands. This is a very popular scheme to produce a smoothed Time Series. Whereas in Single Moving Averages the past observations are weighted equally, Exponential Smoothing assigns exponentially decreasing weights as the observation get older. In the case of moving averages, the weights assigned to the observations are the same and are equal to 1/N. In exponential smoothing, however, there are one or more smoothing parameters to be determined (or estimated) and these choices determine the weights assigned to the observations. This method requires only three items of data: the last periods forecast; the demand for this period and a smoothing parameter, alpha which has a value between 0 and 1.0. The forecast for the next period using exponential smoothing is a smoothing constant, period. The simplest form of exponential smoothing is given by the formulae: (0 1), times the demand in the current period plus (1- smoothing constant) times the forecast for the current

Where is the smoothing factor, and 0 < < 1. In other words, the smoothed statistic st is a simple weighted average of the previous observation xt-1 and the previous smoothed statistic st1. The term smoothing factor applied to here is something of a misnomer, as larger values of actually reduce the level of smoothing. In the limiting case with = 1 the output series is just the same as the original series. Simple exponential smoothing is easily applied, and it produces a smoothed statistic as soon as two observations are available. Values of close to one have less of a smoothing effect and give greater weight to recent changes in the data, while values of closer to zero have a greater smoothing effect and are less responsive to recent changes. There is no formally correct procedure for choosing . Sometimes the statistician's judgment is used to choose an appropriate factor. Alternatively, a statistical technique may be used to optimize the value of . Unlike some other smoothing methods, this technique does not require any minimum number of observation to be made before it begins to produce results. In practice, however, a "good average" will not be achieved until several samples have been averaged together; for example, a

constant signal will take approximately 3/ stages to reach 95% of the actual value. To accurately reconstruct the original signal without information loss all stages of the exponential moving average must also be available, because older samples decay in weight exponentially. Advantages: The major advantage of exponential smoothing methods is that they are simple, intuitive, and easily understood. Generally, exponential smoothing is regarded as an inexpensive technique that gives good forecast in a wide variety of applications. When implemented digitally, exponential smoothing is easier to implement and more efficient to compute as it does not require maintaining a history of previous input data values. Furthermore, there are no sudden effects in the output as occurs with a moving average when an outlying data point passes out of the interval over which you are averaging. With exponential smoothing, the effect of the unusual data fades uniformly. In addition, data storage and computing requirements are minimal, which makes exponential smoothing suitable for real-time application.

Disadvantages The major disadvantage of exponential smoothing methods derives from its basic premise about the model: the level of time series should fluctuate about a constant level or change slowly over time. When the time series takes on an obvious trend, even adaptive exponential smoothing methods will fail to give good forecasting. The exponential smoothing model is, therefore, unable to find subtle nonlinear patterns in the time series data. Obviously, the approximation of linear models to complex real-world problems is not always sufficient. Hence, it is necessary to consider other nonlinear methods to complement the exponential smoothing model.

Analysis:
By using CPI data, we have found out the exponential smoothing forecasts for the following years and then compare them with the actual data.

Assumption:

As we dont have the Forecasting data for any months, we have taken May2010s simple moving average forecasting value to find out the June2010s exponential smoothing forecasting.

We have taken =0.1 as the difference between highest and lowest actual value is not that much.
20092011 Actual Forecas t

Nov Dec Jan Feb Mar Apr

May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr

222.1 4 222.3 3 222.3 8 222.5 1 222.6 1 222.4 9 223.5 9 226.1 1 229.5 5 233.3 1 236.5 3 238.6 8 238.8 9 240.7 5 242.4 8 244.2 8 245.9 7 246.2 4

222.41 222.52 8 222.88 62 223.55 26 224.52 83 225.72 85 227.02 36 228.21 03 229.46 42 230.76 58 232.11 72 233.50 25

May Jun Jul Aug Sep Oct

246.3 9 249.1 1 254.7 2 259.6 6 264.8 5 265.9 4

234.77 63 235.93 76 237.25 49 239.00 14 241.06 72 243.44 55

Table: Exponential Smoothing Forecasting for CPI

A graphical representation of the above calculations is given below:

Figure: Exponential Smoothing forecast for CPI

According to the graphical presentation of the data set, the data trend of the exponential smoothing of the CPI clearly shows an increasing trend in both the actual and forecasted values from May2010 to October2011. We can see from the graph that from the beginning to end always actual CPI value is greater that the forecasting value. So, there is high fluctuation. The differences are particularly larger for the months of June11 to October11.

Actual Vs Forecast:
To find out the effectiveness of the forecasting, we have used Mean Absolute Deviation (MAD) technique for the last six months. By doing so we have found out the following deviation: Month Actua l May Jun 246. 39 249. 11 234.7763 235.9376 11.613 73 13.172 36 11.61373 13.17236 18 Period Forecast Deviation Absolute Deviation

Jul Aug Sep Oct

254. 72 259. 66 264. 85 265. 94

237.2549 239.0014 241.0672 243.4455

17.465 12 20.658 61 23.782 75 22.494 48

17.46512 20.65861 23.78275 22.49448

We can also interpret it by the following graph:

Figure: Actual Vs Exponential Smoothing forecast From the above graph we can see that deviation between actual and forecasting data is lower in May which was 11.62. But from then, it started increasing and reached to 23.78 in September which was not good at all. It means that the forecasting was not that much good if we compare it with the actual figure. But in October, deviation slightly decreased but not enough to say that it was a good forecasting. By using these data we can find out MAD which is as follows:

MAD =

Actual Forecast = 18.19784

This value is lower than the MAD we have found out from the simple average method.