International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-8 Issue-4, November 2019
Identifying Feature Stock Price by Considering
Most Influential Parameters using Prediction
Methods in Indian Stock Exchange
K.Sudhakar, S.Naganjaneyulu
Abstract: Before the evaluation of big data analytics predicting
the optimal share price in the stock market is very difficult, by II. EXAMPLE
applying the big data analytics it is easy to predict frequent
patterns and feature outcomes in any domain. So in this paper we Tcs intrinsic value is 2300 the present current price is 2200
consider the financial domain to predict feature outcomes of then buy signal generated. Means the current price is less than
share prices in the Indian stock exchange. We first gathered the intrinsic value then buy signal is generated otherwise sell
dataset with duration 2011-2016 financial years of TCS signal is generated.
Company, the reason to choose TCS dataset it is a trust based Financial advisors study price charts for price patterns and
company and datasets are available at open access with all
use price data in different calculations to predict future price
parameters. Market price per share is strongly effect by
company’s variable like price earnings, dividend yield, dividend movements. In this paper we consider company‟s specific
per share, earnings per share, book value per share, and return on parameters to predict the future outcomes through linear
equity, after observing the results we conclude that the variables regression.
price earnings, book value per share and firm size are important
determinants of share prices in the Indian stock market. The 2 Motivation
regression model achieved a high R2 (0.94) for the closed price The massive rapid increase in the amount of data and
and book value per share variable and also the model achieved a increasing technological complexities continue to transform
high R2 (0.98) for the closed price and price earnings. the way industries operate and compete. Over the last four
Index Terms: Market price of share, regression, firm specific years, 92 percent of the data in the world has been created as a
variables. result of the creation of 4.5 quintillion bytes of data on a daily
basis, frequently name to as big data. This fast development
I. INTRODUCTION and capacity makes good time for gathering, handling and
investigation of structured and unstructured information.
1 Introduction Organizations utilize information and investigation to
Stock price prediction model is created based on fundamental increase profitable knowledge to advise better business
analysis, technical analysis, machine learning and text mining choices. Businesses that have embraced the utilization of
on articles, news, twitter news; these are the possible ways to enormous information incorporate money related
predict the feature outcomes in stock markets. The company administrations, innovation, promoting and medicinal
performance is identified by using fundamental analysis. services, to give some examples. The selection of big data
Fundamental analysis uses revenues, earnings, future keeps on reclassifying the aggressive scene of industries. An
growth, return on equity, profit margins and other data to expected 91 percent of industries trust those without an
determine a company‟s underlying value and potential for analytics system risk losing a focused edge in the market.
future growth. Financial services, specifically, have generally acquired big
Algorithm for fundamental analysis based on intrinsic data analytics to educate better investments choices with
value reliable returns. In coincidence with big data, algorithmic
Step1: read the intrinsic value, current value if stock. trading utilizes massive historical data with prediction models
Step2: if intrinsic value > current value goto step3. to maximize portfolio returns. Adopting the big data analytics
Step3: buy the stock. to financial services will inevitably gives good prediction
Step4: else sell the stock. results.
Step5: stop. Stock market price prediction is a problem that has the
potential to be worth corers of rupees and is actively
researched by the largest financial corporations in the world.
It is a significant problem; attempts can be made at
approximation using many different machine learning
techniques. The project allows techniques for real-world
Revised Manuscript Received on November 15, 2019 machine learning applications including acquiring and
K.Sudhakar, Assistant Professor G.Pulla Reddy Engineering College analyzing a large data set and using a variety of techniques to
(Autonomous): Kurnool, A.P, India. train the program and predict
S.Naganjaneyulu, Professor of Information Technology Laki Reddy Bali
Reddy Engineering College (Autonomous): Mylavaram, Krishna District,
potential outcomes.
Andhra Pradesh India.
Published By:
Retrieval Number: A9253058119/2019©BEIESP 1 Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A9253.118419 & Sciences Publication
Identifying Feature Stock Price by Considering Most Influential Parameters using Prediction Methods in
Indian Stock Exchange
III. RELATED WORK value per share, dividend payouts, market capitalization
variables are show the impact on Indian stock exchange. This
“The determinants of share cost were started by Collins
paper will focus by leading a far reaching investigation of firm
(1957) for the US market and he recognized profit, net
specific factors, for example, profit per share, profit yield, and
benefit, working income and book an incentive as the
profit per share, book value per share, price earnings ratio and
prominent factors influencing share prices in the United
firm size to decide their degree of effect on the share price of
States”. Irfan and Nishat (2002) distinguished variables
TCS Company.
exerting impact on the share costs in KSE for the period in the
To develop the prediction models, understanding the
vicinity of 1981 and 2000. The review utilized cross-sectional
statistical concepts plays a very important role. Dependent
weighted minimum square regression and broke down the
and independent variables are statistical concepts. Definition
effect of six factors viz. profit yield, payout proportion, size,
of dependent variable and independent variable depend on the
asset development, and procuring instability on share costs,
prediction model. We have adequate information on
Of these the payout proportion, size, use and profit yield rose
statistical concepts in view of learning we build the prediction
as the critical factors influencing the stock exchange prices in
model.
KSE. By observing the above work we infer that, the firm
specific variables significantly affect on market price of
IV. DATA COLLECTION
shares. So many studies reveal price earnings ratios, book
Table 1. Historical Prices of TCS Share
Historical Prices
Company Name: Tata Consultancy Services Ltd.
Exchange : BSE
Open High Close Traded Number Traded
Price Date Price price Low price price value of trades Quantity
12/30/2016 2351.9 2375.9 2351.9 2361.95 98109047 3269 41479
12/29/2016 2304.9 2355 2304.9 2350.75 97507710 2592 41571
12/28/2016 2325 2349 2308.05 2315.8 78697510 2857 33696
12/27/2016 2290 2329 2288.85 2321.85 72082333 2375 31085
12/26/2016 2280 2300 2272.5 2292.1 70389097 2077 30731
12/22/2016 2312.75 2324.7 2286 2309.85 99137747 2334 42989
The information for the review was recovered from the Indian stock R code for linear regression model
exchange site to be specific www.nseindia.com. At present 30
organizations are recorded in the stock exchange and the # load test data sets
organizations fall under the class of Commercial Banks, Investment #identify feature and response
administrations, Insurance organizations, Industrial part, IT variable(s) and values must be numeric
administrations, and tourism organizations. We collected the TCS and numpy arrays
company data with duration of 2011-2016 years. After that we x_train <-
calculated the PE, EPS, and BOOK VALUE etc. input_variables_values_training_datase
ts
V. METHODOLOGY y_train <-
target_variables_values_training_dataset
In this paper we create prediction model by using most s
popular prediction approach that is linear regression. It is x_test<-input_variables_values_test_data
utilized to predict optimal values based on continuous sets
variables. Here, we identify the correlation between x <- cbind(x_train,y_train)
independent and dependent variables by fitting a best line. # Train the model using the training sets
This best fit line is known as regression line and represented and check score
by a linear equation. linear <- lm (y_train ~ data = x)
y= a1 * x + a2. Summary (linear)
Where as
y =dependent variable #Predict Output
a2=intercept predicted= predict(linear x_test)
x= Independent variable
a1=slope.
The coefficients a1 and a2 are inferred in view of limiting the
total of squared distinction of separation between data points
and regression line. We implement this model using R
language.
Published By:
Retrieval Number: A9253058119/2019©BEIESP 2 Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A9253.118419 & Sciences Publication
International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-8 Issue-4, November 2019
1 Return on equity (ROE) =Net Income per Shareholders. And R2=0.987 means we got a strong correlation between PE
and closed price.
By using the line equation we can predict the next optimal
2 Book value per share (BVS) =Total shareholders‟ equity per
outcome.
Number of shares outstanding.
For example y=114.4*19.86+55.35
Y= MPS = 2327.59. Where x= PE of the current year
3 Earnings per share (EPS) = net income per
Table 2. Summary of Variables for TCS share
number of shares.
4 Dividend per share = Dividends paid per Open High low close Trade EPS PE
Price Date Price price price price qty
Number of shares outstanding. 12/30/2016 2351.9 2375.9 2351.9 2361.95 41479 116 20.275
12/29/2016 2304.9 2355 2304.9 2350.75 41571 116 19.86983
5 Dividend yield (DY) = Dividend per
12/28/2016 2325 2349 2308.05 2315.8 33696 116 19.89698
share per price per share. 12/27/2016 2290 2329 2288.85 2321.85 31085 116 19.73147
12/26/2016 2280 2300 2272.5 2292.1 30731 116 19.59052
6 Price earnings (PE) = Stock price per EPS.
12/23/2016 2309 2315.7 2280 2290.2 28276 116 19.65517
12/22/2016 2312.75 2324.7 2286 2309.85 42989 116 19.7069
7 Debt to total assets (DA) =Total debt per Total Assets 12/21/2016 2345 2345 2305 2312.75 31711 116 19.87069
The types of variables used in our model are Actually in the retrieved dataset there is no EPS and PE
Independent Variables values, we calculated the EPS, PE, BVS, etc by using
Control Variable standard formulas. In the above dataset the PE value on
Firm size Market capitalization (MACP) = Market price of 29/12/2016 is 19.86, that is used in our model to get next day
share *number of shares outstanding. predicted value, after substitution the values we got 2327.59.
Dependent Variable The Closed price on 30/12/2016 is 2361.95. The outcome
Market price of share (MPS) = Closing share price. (or) predicted value is 2327.59 which is close to actual closed
Linear regression identify the correlation and R2 between the price. If we consider 30/12/2016 PE value then we can
independent and dependent variables, based on R2 we can get identify the predicted value of 1/1/2017 and so on. Next we
strongest correlation between variables. find the correlation between close price and BVS variables.
The following figure shows the relation between close price
and BVS.
Fig no 1: Estimation model or prediction model
VI. RESULTS
Fig 3: relation between close price and BVS (book value
per share)
The x-axis represents the BVS (book value per share). The
y-axis represents the close price. There is a strong correlation
between BVS and close price, we identify based on R2, and
the R2 is 0.944 means 94%. This means prediction accuracy is
94%.
Fig 2: shows the relation between PE(X- axis) and closed Next we identify the relation between traded volume and
price (Y-axis) close price. The following figure shows the relation between
The figure 2 represents the linear regression equation that is traded volume and close
obtained by giving input data to linear regression algorithm. price.
The equation y=114.4x+55.35
Published By:
Retrieval Number: A9253058119/2019©BEIESP 3 Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A9253.118419 & Sciences Publication
Identifying Feature Stock Price by Considering Most Influential Parameters using Prediction Methods in
Indian Stock Exchange
AUTHORS PROFILE
K.Sudhakar, is currently pursuing the Ph.D
degree with Jawaharlal Nehru Technological
University, Kakinada, and Andhra Pradesh, India
and working as an Assistant Professor in G.Pulla
Reddy Engineering College (Autonomous):
Kurnool, A.P, India. His research interests
include Big Data analytics, cloud computing,
Machine Learning, Data Mining.
Dr. S. Naganjaneyulu, working as a professor of
Information Technology in Laki Reddy Bali
Reddy Engineering College (Autonomous):
Mylavaram, Krishna District, Andhra Pradesh
Fig 4: relation between volume and close price. India. His Research interests include Data Mining,
We achieved the low R2 that is 0.075 means the prediction Big Data.
accuracy is 7.5%. So we concluded the close price is not
depending on traded volume. The trend line also shows the
same. If too much volume is increased the price value is
decreased.
VII. CONCLUSIONS
The foremost intension of this work is to inspect the
determinants of market price of TCS share. The research
mostly try to found a relationship between market price of
shares (MPS) and additional variables specifically price
earnings (PE), book value of share (BVS), earnings per share
(EPS), traded volume, return on equity (ROE), dividend per
share (DPS), dividend yield (DY) , debt to total asset (DA)
and firm size (MCAP).
The factual results show effective and notable correlation
between PE, BVS, and traded volume recommended that
these variables perform as influential determinants in making
the market price of shares.
The empirical results also shows the notable negative
correlation was identified between traded volume, and closed
price. This suggests that traded volume is not an optimal
variable to predict the closed price. Hence traded volumes
cannot attract the investors. Closed price is closed related to
the market price of share, here we consider predicted closed
price as a feature market price of TCS share.
Recommendations
Investors can expand their portfolio simply by observing
the PE ratios, book value per share, market capitalization of
the company and ROE variables.
Limitations
Here we are not consider the Macro-Economic factors like
monsoon, inflation, interest rates, GDP, etc and also not
consider the articles, magazines, tweets. In India Stock
markets are highly affect by various kinds of news, survey
reports, ratings, and latest demonetization effect.
REFERENCES
1. Random walks in stock market prices [J]. Financial Analysts
Journal,1965: 55-59. Fama E F.
2. Baker M, Wurgler J. Investor sentiment in the stock market[J]. 2007.
3. Weron, R.: „Modeling and forecasting electricity load prices: a
statistical approach‟
4. Parameters Optimization in Algorithm Trading using Hadoop
MapReduce Meibin Chen, Yifan Wu
Published By:
Retrieval Number: A9253058119/2019©BEIESP 4 Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A9253.118419 & Sciences Publication
