(IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 8, No.7, 2017 388 | Page www.ijacsa.thesai.org Financial Market Prediction using Google Trends Farrukh Ahmed Research Student, Department of Computer Science & Software Engineering NED University of Engineering & Technology, Karachi, Pakistan, 75270 Dr. Raheela Asif Assistant Professor, Department of Computer Science & Software Engineering NED University of Engineering & Technology, Karachi, Pakistan, 75270 Dr. Saman Hina Assistant Professor, Department of Computer Science & Software Engineering NED University of Engineering & Technology, Karachi, Pakistan, 75270 Muhammad Muzammil Research Student, Department of Computer Science & Software Engineering NED University of Engineering & Technology, Karachi, Pakistan, 75270 Abstract—Financial decisions are among the most significant life-changing decisions that individuals make. There is a strong correlation between financial decision making and human behavior. In this research the relationship between what people think and how stock market moves is investigated. The data from 2010 to 2015 of some of business, political and financial events which directly impact the local stock market in Pakistan is analyzed. The data was collected from search engine Google via Google trends. The association between internet searches regarding the political or business events and how the subsequent stock market moves is established. It was found that increase in search of these topics may lead to stock market fall or rise. The overall objective of this research is to predict Karachi Stock Exchange (now known as Pakistan stock exchange) 100 index by quantifying the semantics of international market. In addition to that, the relation between what an individual thinks while searching on Google which affects the local market is also investigated. The collected data has been mined by Multiclass Neural Network and Multiclass Decision Trees. The result shows that Multiclass Decision Trees performed best with an accuracy of 94%. Keywords—Google trends; financial market; stock market; Karachi stock market; multiclass neural network; multiclass decision trees I. INTRODUCTION We are living in a world where data is generated from all domains of life. For example, from every social media interaction do, from every computer, every mobile, every sensor and now even from watches and other wearable gadgets. The real question is how we can convert this data into meaningful information for decision making such as to predict stock market behavior. Stock market prediction is a domain of challenging factors which is based on many important aspects and collective thinking of the financial experts. Stock Market data can be acquired from different sources. Its impact has generated considerable scientific attention due to its complexity and size. Despite its huge size, such data sets capture only the final action taken at the end of a decision- making process. No insight is provided into earlier stages of this process, where traders may use this information to determine what consequences of various actions and factors may be. The aim of this study is to predict the behavior of local stock market in Pakistan based on available historical data and International market factors such as International Gold Rates, US dollar Rates, International stock markets and foreign exchange reserves etc. For today's world, data gathering frequently comprises of seeking on the web sources. Few years back Google has given access to cumulative information on the volume of queries for different search terms and how these volumes change over time, via publicly available service named Google Trends [1]. The gathered data was pre- processed using data cleaning and data filtering. The preprocessed data was than analyzed using Machine learning algorithms. II. LITERATURE REVIEW The advent of Internet has seen people have used it as a main source of Information gathering and search engines like Google have become a gateway to this information. This fine grained data available on internet has opened up new options for researchers. Studies have found that large volumes of search queries for a specific word can linked up to real life events, such as forecasting the housing prices and sales [2]; popularity of films, games, and music on their release [3]; unemployment rate [4]; This openly available digital data also help researchers to find how the stock market moves, a recent study has found that increase in search volumes of some topics tends to precede stock market falls [5]. Some researchers have successfully found the relationship between behavior of people through social media (like twitter) and prediction of the stock market [6]. Karachi Stock Market (KSM) is one of the top 10 markets in the world. There are dozens of factors which impacts stock exchange directly or indirectly. That’s why this research was intended to work on unique factors which impacts stock exchange. The objective of this research was to predict complex behavior of Karachi stock market (KSM) using historical data of KSM in combination with International economic factors such as US Dollar rates, gold prices, Oil