Final Report on Aerosol

download Final Report on Aerosol

of 23

Transcript of Final Report on Aerosol

  • 8/18/2019 Final Report on Aerosol

    1/23

    Analysis of Forecasting Demand of Aerosol

    2.3.1 Forecasting

    Forecasting is the process of making statements about events whose actual outcomes

    (typically) have not yet been observed. A commonplace example might be estimation for 

    some variable of interest at some specified future date. Prediction is a similar, but more

    general term. Forecasting is an important tool for the future of demand condition. his is

    a prediction of future events used for planning purposes. Forecasting is needed to aid in

    determining what resources are needed, scheduling existing resources and ac!uiring

    additional resources.

    2.3.2 Types of Forecasting

    "ainly, the forecasting can be classified as#

    $ualitative methods

    $uantitative methods

    Qualitative Methods# $ualitative forecasting techni!ues generally employ the %udgment

    of experts in the appropriate field to generate forecasts. A key advantage of these

     procedures is that they can be applied in situations where historical data are simply notavailable. "oreover, even when historical data are available, significant changes in

    environmental conditions affecting the relevant time series may make the use of past data

    irrelevant and !uestionable in forecasting future values of the time series.

    Quantitative Methods# include causal methods and time&series analysis. 'ausal methods

    are historical data on independent variables, such as promotional campaigns, economic

    conditions and competitors actions to predict demand or supply. ime series analysis is a

    statistical approach that relies heavily on historical demand data to pro%ect the future sie

    of demand or supply and recognies trends and seasonal patterns.

    o make our analysis we have followed the following techni!ues of forecasting.

    *) +imple "oving Average

    ) +ingle exponential +moothing

    -) +imple inear /egression

    Page 0 *

    http://en.wikipedia.org/wiki/Estimationhttp://en.wikipedia.org/wiki/Predictionhttp://en.wikipedia.org/wiki/Predictionhttp://en.wikipedia.org/wiki/Estimation

  • 8/18/2019 Final Report on Aerosol

    2/23

    Analysis of Forecasting Demand of Aerosol

    1) Simple Moving Average

    he simple moving average forecast uses a number of the recent actual data values in

    generating a forecast. he moving average forecast can be computed using the following

    e!uation#

     F t =

    ∑t =1

    n

     A t −i

    n

    ¿ A t −n+…+ A t −2+ A t −1

    n

    1here,   F t  2 Forecast for the time period t

     A t −1  2 Actual value in period t −1

     MA t   2n period moving average

     32 number of periods (data points) in the moving average

    2) Single Exponential Smooting

    +ingle 4xponential +moothing largely overcomes the limitations of moving average

    models. 4ach new forecast is based on the previous forecast plus a percentage of the

    difference between that forecast and the actual value of the series at that point.

     F t = F t −1+α ( A t −1− F t −1)

    1here,   F t   2 Forecast for period t

     F t −1 2 Forecast for previous period

    α2 +moothing constant

     A t −1  2 Actual demand for previous period

    Page 0

  • 8/18/2019 Final Report on Aerosol

    3/23

    Analysis of Forecasting Demand of Aerosol

    3) Simple !inear "egression

    his method involves linear relationship between two variables. he ob%ect in linear 

    regression is to obtain an e!uation of a straight line that minimies the sum of s!uared

    vertical deviations of data points from the line. he e!uation is#

     y=a+bx

    1here, a 2 constant and b 2 slope of the line

    And,

    a=´ y−b ´ x

    b=n∑ xy−∑  x∑  yn∑ x2−(∑  x2 )

    2.3.2 Acc#racy Test

    After forecasting, applying different techni!ues, we have found out the forecasting

    accuracy and the appropriate techni!ues for forecasting the sales of the product. o

    calculate the forecasting accuracy, we have used the following techni!ues#

    "ean absolute deviation ("A5) "ean s!uared error ("+4)

    "A5 is the average absolute error, "+4 is the average of s!uared error and "AP4 is the

    average percent error. he formulas used to compute "A5, "+4 are given below#

    ∑ ¿ Actual− Forecast ∨¿n

     MAD=¿

    Page 0 -

  • 8/18/2019 Final Report on Aerosol

    4/23

    Analysis of Forecasting Demand of Aerosol

     MSE=∑ ( Actual− Forecast )2

    n−1

    "+4 is similar to the variance of a random sample6 however, it is more sensitive to a fewlarge errors than "A5. 'onse!uently, "A5, the average of the absolute discrepancies

     between the actual and fitted values in a given time series is often preferred. 7f a model

    fits the past time&series data perfectly, the "A5 value would be ero. As the fit worsens,

    the value of "A5 increases. 7n other words, a small value of "A5 is desirable. 7n

    addition, when forecast errors are normally distributed, an estimate of the standard

    deviation of the forecast error is given by *.8 times "A5. 1e also considered "AP4

    test. he advantage of this measure of accuracy is that "AP4 is not dependent on the

    magnitude of the values of demand.

    2.1.$ Data Analysis Tool

    7n order to get the forecasting output of aerosol, we have used "icrosoft 1ord and

    "icrosoft 4xcel.

    2.2 !imitations

    his study has been done based on the data of sales volume and production from 9anuary

    :*; to used it for the remaining !uarters. 3ot enough

    statistical data has allowed little space for in depth analysis > thus to some extent might

    have hampered the accuracy.

    Page 0 ;

  • 8/18/2019 Final Report on Aerosol

    5/23

    Analysis of Forecasting Demand of Aerosol

    Chapter - 03

    1.%A&A!'S(S

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    6/23

    Analysis of Forecasting Demand of Aerosol

    3.1 Simple Moving Average

    he simple moving average forecast uses a number of the recent actual data values in

    generating a forecast. he following figures shows the - simple moving average on the

     basis of sales data which is the actual demand for :*-&:*8, and we figure the forecast

    of these three years for different +CD 8:ml, -8:ml, ;8:ml, and ?::ml.

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    7/23

    Analysis of Forecasting Demand of Aerosol

    0

    5000

    10000

    15000

    20000

    2500030000

    35000

    40000

    2% ml

    250 ml 2015 Actual Demand

    250 ml 2015 Simple Moin! Aera!e

    05000

    1000015000200002500030000350004000045000

    3% ml

    350 ml 2015 Actual Demand

    350 ml 2015 Simple Moin! Aera!e

    Page 0 @

  • 8/18/2019 Final Report on Aerosol

    8/23

    Analysis of Forecasting Demand of Aerosol

    0

    50000

    100000

    150000

    200000250000

    *+ ml

    475 ml 2015 Actual Demand

    475 ml 2015 Simple Moin! Aera!e

    0

    20000

    40000

    60000

    80000

    100000

    120000

    ,%% ml

    800 ml 2015 Actual Demand 800 ml 2015 Simple Moin! Aera!e

    Page 0 ?

  • 8/18/2019 Final Report on Aerosol

    9/23

    Analysis of Forecasting Demand of Aerosol

    3.2 Single Exponential Smooting

    =ere we took α =0.3  to do our analysis because when we divide maximum value with

    minimum value it comes out more than .

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    10/23

    Analysis of Forecasting Demand of Aerosol

    0

    10000

    20000

    30000

    40000

    50000

    60000

    2%ml

    250 ml 2015 Actual Demand250 ml 2015 Sin!le eponential Smoothin!

    0

    10000

    20000

    30000

    40000

    50000

    60000

    3% ml

    350 ml 2015 Actual Demand350 ml 2015 Sin!le eponential Smoothin!

    Page 0 *:

  • 8/18/2019 Final Report on Aerosol

    11/23

    Analysis of Forecasting Demand of Aerosol

    0

    50000

    100000

    150000

    200000

    250000

    *+ ml

    475 M 2015 Actual Demand

    475 M 2015 Sin!le eponential Smoothin!

    0

    20000

    40000

    60000

    80000100000

    120000

    140000

    ,%% ml

    800 ml 2015 Actual Demand

    800 ml 2015 Sin!le eponential Smoothin!

    Page 0 **

  • 8/18/2019 Final Report on Aerosol

    12/23

    Analysis of Forecasting Demand of Aerosol

    3.3 Simple !inear "egression

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    13/23

    Analysis of Forecasting Demand of Aerosol

    -

    5,000

    10,000

    15,000

    20,000

    25,00030,000

    35,000

    40,000

    250 ml

    Actual Demand Simple inear .e!re''ion

    -

    5,000

    10,000

    15,000

    20,000

    25,00030,000

    35,000

    40,000

    350 ml

    Actual Demand Simple inear .e!re''ion

    Page 0 *-

  • 8/18/2019 Final Report on Aerosol

    14/23

    Analysis of Forecasting Demand of Aerosol

    -20,00040,00060,00080,000

    100,000120,000140,000160,000

    475 ml

    Actual Demand Simple inear .e!re''ion

    -

    10,000

    20,000

    30,00040,000

    50,000

    60,000

    70,000

    80,000

    800 ml

    Actual Demand Simple inear .e!re''ion

    Page 0 *;

  • 8/18/2019 Final Report on Aerosol

    15/23

    Analysis of Forecasting Demand of Aerosol

    3.* Forecasting Di-erences

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    16/23

    Analysis of Forecasting Demand of Aerosol

    0

    10000

    20000

    30000

    40000

    50000

    60000

    70000

    80000

    Forecasting Di-erences et/een Di-erent Metods

    $oreca'tin! o/ AC Simple Moin! Aera!e

    Sin!le +ponential Smoothin! Simple inear .e!re''ion

    Page 0 *B

  • 8/18/2019 Final Report on Aerosol

    17/23

    Analysis of Forecasting Demand of Aerosol

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    18/23

    Analysis of Forecasting Demand of Aerosol

    0

    10000

    20000

    30000

    40000

    50000

    60000

    70000

    Forecasting Di-erences et/een Di-erent Metods

    $oreca'tin! o/ AC Simple Moin! Aera!e

    Sin!le +ponential Smoothin! Simple inear .e!re''ion

    Page 0 *?

  • 8/18/2019 Final Report on Aerosol

    19/23

    Analysis of Forecasting Demand of Aerosol

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    20/23

    Analysis of Forecasting Demand of Aerosol

    0

    50000

    100000

    150000

    200000

    250000

    Forecasting Di-erences et/een Di-erent Metods

    $oreca'tin! o/ AC Simple Moin! Aera!e

    Sin!le +ponential Smoothin! Simple inear .e!re''ion

    Page 0 :

  • 8/18/2019 Final Report on Aerosol

    21/23

    Analysis of Forecasting Demand of Aerosol

    5ata from 5ecember :*;&

  • 8/18/2019 Final Report on Aerosol

    22/23

    Analysis of Forecasting Demand of Aerosol

    0

    20000

    40000

    60000

    80000

    100000

    120000

    140000

    Forecasting Di-erences et/een Di-erent Metods

    $oreca'tin! o/ AC Simple Moin! Aera!e

    Sin!le +ponential Smoothin! Simple inear .e!re''ion

    Page 0

  • 8/18/2019 Final Report on Aerosol

    23/23

    Analysis of Forecasting Demand of Aerosol

    3. Acc#racy Test

    S 250ml 350ml 475ml 800ml

    +imple

    "oving

    Average

    MAD 2802 3674 11163 7502

    MS' 1212707 1778276 128008715 82131571

    +ingle

    4xponential

    +moothing

    MAD 150 8305 26204 1868

    MS'

    112772476 10847187 10716066 5518441

    +imple

    inear 

    /egression

    MAD 538 63 3823 2774

    MS' 383524 635885 1351211 1010340

    7n the above chart we can see that simple linear regression is giving out better figures in

    terms of comparison between it, simple moving average and single exponential

    smoothing for forecasting. As we know the lower the figure the better the model. hus for 

    further forecasting we can use the simple linear regression.

     

    Page 0 -