Machine successfully trade securities for greater gain thus beating



Improvements in the
machine learning algorithms have immensely helped the traders in evaluating and
implementing the potential predictive algorithms to get optimum profit in
financial security markets. Machine learning is segregated into two classes, the
first class is supervised learning, in which output corresponding to that
feature set. This means that the algorithm is given features and outputs for a
particular dataset (training data), and must apply what it “learns” from this
dataset to predict the outputs (labels) for another dataset (test data).
Unsupervised learning, on the other hand, consists of examples where the
feature set is unlabeled. The algorithms generally try to cluster the data into
distinct groups. Supervised learning can be further broken down into
classification and regression problems. In classification problems there are a
set number of outputs that a feature set can be labeled as, whereas the output
can take on continuous values in regression problems.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now



Due to the
unpredictable behavior of stock market many experienced investors in Pakistan
feared for making bigger investments in stock market as they do not suitably
know as to which stocks to buy, sell or hold to reap greater margin returns and
rather prefer to investment in real estate businesses. In order to accelerate
the growth of this potential field there is a dire need of automated predicting
model on short term basis which can be used for making more money off
securities rather than traditional buy and hold strategy also to lessen the
risk involved in doing such investments. Most of the models available and use
today are based on predicting long term values which includes greater margin of
error due to rapid fluctuation of stocks based on world events.



Objective of this
research work is to develop a stock prediction model using Business
Intelligence tools which can successfully trade securities for greater gain
thus beating the conventional buy-and-hold market strategies. The proposed
model will incorporate technical features as well as market and environmental
sentiments for predicting the share prices with greater accuracy. The model
will be evaluated by classification and profitability as compared to
buy-and-hold strategy.


The performance
evaluation of the build model will be carried out using following performance
evaluation methodologies:·        
Classification Percentage

False Positive percentage

Profit versus buy-and-hold Strategy.



Technical objectives
of the model will be implemented using R programming language. The system
should be able to access the historical prices and must calculate the estimated
prices stock based on historical data and current events. The model must
understand data properties and provide s the instantaneous visualizations of
the share prices thru charts and graphs.



The model will be
implemented using two different approaches; one using basic technical
indicators which incorporate Moving Averages (MA), Moving Exponential Averages
(EMA), Bollinger Bands and other using Time Series ARIMA and HOLT WINTER model.
Results of the both the approaches will be analyzed to find prediction accuracy
of both.



research study will answer the following research questions:


Q1) Prediction of
future stock prices of Cement Sector of Pakistan

Q2) Produce effective
patterns from past data for analysis.

Q3) Analyze the
non-financial factors and news sentiments affecting the stock prices in

Q4) Bring in novice
and feared investors into the market.


of research:

Forecasting of stock market scripts may provide as an
early recommendation for short term investors these market predictions may
serve as early distress warnings for long term stake holders also. Many stock
forecasting studies focus on macroeconomic indicators such as CPI and GDP, to
train and develop the prediction model but in order to predict the market
movement on short term basis daily data of all the major macroeconomic
indicators is impossible to attain timely making it difficult on short term
basis. In this research study we propose a method that directly uses prices
data to predict stock price direction. An extensive empirical study of the
proposed method will be presented on cement sector of Pakistan Stock Exchange (PSX).The
research study on this topic will be beneficial in identifying the various
aspects related to the profitability of cement industry in Pakistan.

of the Research:

Abundance of Forecasting Models

Abundance of trading platforms/programming

Abundance of technical indicators

Use of financial data relating to stock
market indices