BIRLA INSTITUTE OF TECHNOLOGY AND
SCIENCE, PILANI. Hyderabad campus.
End Semester Report
Submitted in partial fulfilment of the requirements of
Course Name: PHA G616 Pharmaceutical Administration &
Title: Forecasting by Linear Regression Analysis
Report Submitted by: Suvarna Kale.
Instructor-in-charge: Dr. Akash Chaurasiya.
TOC o “1-3” h z u Course Name: PHA G616 Pharmaceutical Administration ; PAGEREF _Toc527895026 h 1Management (PAM) PAGEREF _Toc527895027 h 1Evaluator’s remarks: PAGEREF _Toc527895028 h 1Forecasting PAGEREF _Toc527895029 h 3Types of Forecasting PAGEREF _Toc527895030 h 3•Short range PAGEREF _Toc527895031 h 3•Long range PAGEREF _Toc527895032 h 3Forecasting Process: PAGEREF _Toc527895033 h 3Methods of Forecasting: PAGEREF _Toc527895034 h 4Linear Regression Analysis PAGEREF _Toc527895035 h 5Variables role: PAGEREF _Toc527895036 h 5Dependent variable: PAGEREF _Toc527895037 h 5Independent Variable: PAGEREF _Toc527895038 h 5?Linear Regression Function PAGEREF _Toc527895039 h 5REGRESSION MODELS PAGEREF _Toc527895040 h 6Case Study: A Linear Regression Approach to Prediction of Stock Market Trading Volume. PAGEREF _Toc527895041 h 6ABSTRACT PAGEREF _Toc527895042 h 6INTRODUCTION PAGEREF _Toc527895043 h 6•Scatter plot PAGEREF _Toc527895044 h 7Results after applying Regression Formula: PAGEREF _Toc527895045 h 9Conclusion PAGEREF _Toc527895046 h 10References: PAGEREF _Toc527895047 h 11
Forecasting is define as the process of making predictions of the future based on past and present data and mostly by analysis of trends (recently observed trend data).
Most of the Business Enterprises uses Forecasting as an integral part of business.
Fig 1. Forecasting.
Types of ForecastingForecasting may be:
Short range: Forecasting for an hour, day, week or month.
Long range: Forecasting for next six months, next year, the next five years, or the life of product or service.
Forecasting Process: There are various methods and Techniques by which the forecasting is done depending on the condition:
Methods of Forecasting:
Linear Regression AnalysisLinear regression: It is a mathematical regression technique which relates one variable ie.independent variable to that of another, the dependent variable which gives a straight line.
General form of a linear equation is as following:
y = a+bx
y = the dependent variable
a = the intercept
b = slope of the line
x = the independent variable.
Fig.Linear Regression Graph.
Variables role:Dependent variable:This is the variable whose values we want to explain, estimate or forecast.
Its value is dependent on something else.
It is denoted by y.
Independent Variable:This is the variable that explains the other one/ to predict the dependent variable
Its value are independent
It is denoted by x.
Linear Regression Function:
y=a + bx
Slope(b) of the Regression line:
b=r* STD deviation of y/std deviation of x
Y-intercept(a) of Regression line:
a=mean of y – b * mean of x
When there is use of more than one independent variable in regression analysis then it is called as Multiple Regression analysis.
Stock Marketing Case Study: Prediction of Stock Market Trading Volume by Linear Regression Approach.
Stock market is the business were people invest their money for gaining profit, but there is risk of fluctuations as per the market. Predicting the daily stock is a very much great challenge for the investors and stockholder. So Forecasting by Linear regression helps the investor and the stockholders to invest the shares very safely and confidently. In this Paper by applying the above method we get similar and good performance in compare to the real volume so that the investor can invest confidently based on it.
INTRODUCTIONDue to financial benefits and its low risk is a growing topic in research, predicting the stock market due to its importance and popularity among the masses and also small and large companies is very much important.
There is flucation effect on the behaviour of the people in terms of investment, capital savings, stock price or the decrease or increase risk.
So in general there are various methods to predict the stock market behaviour for buying the share at cheapest price and selling it at highest price. So, it is very useful to choose an appropriate method for forecasting.
Regression is used to predict a numerical value.
The result can be extended by adding new information where the Target values have been defined already.
Establishes values between Predictor and Target values.
Scatter Plot of :
(Average) parameter is the mean of the prices of Open, Low, High and close to predict
There is Relationship between:
Trading volume (Volume) as the dependent variable.
Average price per share (Average) as the independent variable.
The R-squared 0.358 simply determines that the 2 variables used for determining the orientation of trend line was 35.5%.
The relationship between the independent parameters:
Open Close Low High
Low 0.975 0.986 – –
High 0.989 0.976 0.985 –
Volume -0.383 -.0429 -0.425 -0.391
Linear regression was applied to the data by using data analysis and the summary of data is as follows:
Regression value by applying the regression analysis by using Regression Statistics.
Multiple R 0.599
R Square 0.358
Adjusted R Square 0.347
Standard Error 285577
Regression analysis: The Std error:
It measures the dispersed or scatter of the observed value along the line of regression for the value of X.
Confidence and prediction interval can be calculated by using the STD error.
The STD error is equal to 285577 which is the error between the predicted and real value and is calculated by using the degree of freedom, sum and mean of square.
ANOVA df SS MS
Regression 1 3E+12 3E+12
Residual 57 5E+12 8E+10
Total 58 7E+12 Values of coefficient obtained after linear regression:
Coefficient Std error
Intercept 4675513 697440
Average -106938 18953
Results after applying Regression Formula:By obtaining the values of slope, coefficients, error and intercept and then applying the linear regression on data for predicting the trading volume which is unknown parameter to the real volume by using formula.
Following table represents the predicted volume and the real volume of the average price:
Date Avg Price Predicted Volume Real Volume
28/6/13 33.13 965675 1081200
27/6/13 32.96 963498 801800
26/6/13 32.49 1019301 835100
25/6/13 32.34 1042679 196700
24/6/13 32.87 985747 1017800
21/6/13 32.94 964849 1196100
20/6/13 33.11 955342 1156600
18/6/13 33.53 921192 1794100
17/6/13 33.52 919738 2512100
ss14/6/13 33.52 918908 645000
7/6/13 35.88 788848 656100
The predicted Trading volume is almost very similar to the real values.
The similarity of about 61% was observed between the real and the predicted value observed.
So, The Linear Regression is one of the most valuable and useful technique for prediction of stock market trading volume.
References:Farhad Soleimanian Gharehchopogh1, Tahmineh Haddadi Bonab2 And Seyyed Reza Khaze,Linear Regression: Approach To Prediction Of Stock Market Trading Volume: A Case Study, International Journal Of Managing Value And Supply Chains (IJMVSC) Vol.4, No. 3, September 2013.