In my assignment Let me forecast the 3rd and the fourth quarter income of Consolidated Edison Firm for the year 1996. You can actually main areas are electrical power, gas and steam offering. In the case of just about every company it is important to prediction the future profits to be able to calculate the company’s expected profits. This provides the situation in this instance as well, thus i must do my personal job since perfect as I will.
I got yesteryear eleven years data, from which I can analyse the whole condition and that we can use to predict for the future. To make the forecast more accurate I could use the real quarterly profits.
Quarterly profits for Consolidated Edison Business
($ million), 1985-1995
Year March 23. June 35. September 35. December thirty-one.
1985 1441 1209 1526 1321
1986 1414 1187 1411 1185
1987 1284 1125 1493 1192
1988 1327 1102 1469 1213
1989 1387 1218 1575 1371
1990 1494 1263 1613 1369
1991 1479 1330 1720 1344
1992 1456 1280 1717 1480
1993 1586 1396 1850 1483
year 1994 1697 1392 1822 1461
1995 1669 1460 1880 1528
1996 1867 1540
Source: The Value Series Investment Survey (New York: Value Range, 1990, 93, 1996) g. 170.
There are numerous different methods, which can be used by forecasters. Just for this case Let me test the nave, the moving uses, the rapid smoothing, the double shifting averages, the deseasonalisation, the linear regression and the rapid regression models.
After having conducted the procedures, the forecaster’s task is to assess the models. This is not an easy task since there are a lot of measures, based on which the person has to make a decision.
The evaluate coefficients evaluation the difference involving the observed as well as the forecasted principles, which then utilized for comparison.
These types of measures are as follows:
MSE: This is the imply squared problem, which amount and sq . all of the problems and consider their common.
MAD: This is actually the mean complete deviation, which in turn sum the absolute errors and take all their averages.
MAPE: This is the suggest average percentage error, which usually shows the in percentages.
As I stated, all of these procedures test the errors, so when the ideals of procedures are the tiniest in a approach, that technique seems to be one of the most accurate one particular.
Now, I will conduct the several methods 1 by 1.
The 1st technique is the naive strategy. The substance of this strategy is that by using the value of the present period as the forecast for the next period. This model is definitely rarely the best one since it does not take the seasonality and the economic adjustments into consideration. (Table I)
Another method I’ve conducted is the moving uses. This technique uses several past time periods while the forecast for the next period. I proportioned three and four sectors to get the feasible best a single, but it offers turned up which the three quarter one has overestimated, while the four quarter you have underestimated the values a lttle bit. From the chart we can see which the four one fourth moving average method does not take the seasonality into consideration, consequently it calculates only typical values. (Table 2)
Following the moving uses procedure I actually conducted the exponential smoothing method, which will uses a measured average of past time series values to have a smoothed prediction. This model reduces the effects of earlier data which way makes more accurate forecasts for the future. I used 3 different weight loads, the value of 0. 2 as well as the 0. 4 and the zero. 8. Among them the model measured by zero. 2 was the most accurate one. (Table 3)
The double going average model is a better variation of the moving averages models. Even though a better end result was wished from this strategy I must admit the result was worse compared to the previous ones’. It is noticed in the chart that this method is continually overestimating. It can be related to the wrong version building. It might be interesting to try the several quarter a single as well. (Table 4)
It is known that usually the best procedure is the deseasonalisation technique, because this method splits the constituents of the time series up in to parts and analysed independently. After, the constituents are remanufactured and the outlook is made. (Table 5)
The regression types (linear and exponential) utilize the built in regression of Excel to prediction the values. The different types are needed because the beliefs of data could possibly be fit to a straight range or to an exponential curve. To be able to find the possible finest results I need to conduct them all. (Table six, 7)
Having conducted the several procedures it is time to compare these to each other.
MSE MAD MAPE
Naive prediction 1 period ahead 74644. 33 254. 56 18. 45%
Shifting Average 3quarter 37631. 80 168. 74 11. 51%
Moving Typical 4quarter 24499. 71 136. 20 9. 23%
Exponential Smoothing a=0. 2 29485. 20 146. 25 9. 98%
Rapid Smoothing a=0. 4 33731. 96 one hundred sixty. 62 15. 97%
Rapid Smoothing a=0. 8 54578. 95 213. 13 13. 59%
Deseasonalisation 7010. 60 62. 04 4. 43%
Double Going Average 3quarter 71216. 02 232. 20 15. 89%
Linear Regression 23354. 74 134. forty-eight 9. 33%
Exponential Regression 23343. 68 132. 57 9. 12%
From the desk it is evidently seen which the values of error conditions are the most compact for the deseasonalisation unit in all from the three procedures of accuracy. In Graph 2 it is seen the forecasted beliefs closely suit to the past data. This means that that I have to forecast with this method to be the most effective.
The deseasonalization style operates with splitting the time series in components, which are the trend, the cyclical, the seasonal as well as the irregular component.
In time series the trend element is the long-term component that represents the expansion or drop in the series over a period of period. In the case of the Consolidated Edison Company, this kind of trend impact is a constant growth, which includes started seeing that 1985. This kind of trend impact can be associated with the changes inside the economy- pumpiing and the constantly growing intake.
The cyclical component may be the wavelike changing around the tendency. Any frequent pattern over or under the trend series might be linked to the effect of cyclical aspect. In this case this component firmly affected the season 1985-1987, but after this short time it reduced and features much weakened affect for the revenues.
The seasonal aspect refers to a pattern of change that repeats alone year after year. This seasonal component causes the fluctuation of revenues inside the different sectors. These improvements can be considered as the effect of weather alterations and some other regular within a year time.
The irregular component is definitely the measure of variability of the time sequence soon after the other components after the other pieces have been removed. This aspect can decide the capricious and unexpected factors, which usually causes doubt for the forecast. Within my case this kind of component can be filtered away by the averaging procedure.
Since I have previously presented the model which i found the best I must complete the original task, the outlook.
Based on the pc output the trend values pertaining to the third and the fourth one fourth are 1779 and 1792. To reach the forecasts I must multiple these types of trend principles by the periodic components, which are 1 . 138 and 0. 929.
Prior to I inform the result of prediction I want to review these seasons indexes.
The value of 1 . 138 means that in each third quarter the revenues happen to be above the craze line by simply 13. 8% on an normal. The value of 0. 929 signifies that the earnings in every fourth one fourth are several. 1% below the trend collection on an common.
Finally the reached the desired info is 2024 intended for the third quarter and 1666 for your fourth period. These values mean that the profits are expected to become $2. 024 billion inside the third quarter of mil novecentos e noventa e seis, while the earnings in the next quarter are expected to be $1. 666 billion in mil novecentos e noventa e seis.
Economics Works
