Sie sind auf Seite 1von 29

RNS INSTITUTE OF TECHNOLOGY

BENGALURU - 98

DEPARTMENT OF INFORMATION SCIENCE &


ENGINEERING

Presentation on
Internship project

Predicting the Bitcoin Market Capital


using Random Forest in Machine
Learning

Under the guidance of


Sushma P R Mrs. Geetanjali C N
USN: 1RN16IS107 Prof, Dept of ISE, RNSIT
AGENDA

 Abstract
 About the Company
 Introduction
 Literature Survey
 Requirements
 System Design
 Detailed Design
 Implementation
 Conclusion and Future Enhancements
 References
 Q &A
ABSTRACT
 Bitcoin Market Capital prediction is done using Random Forest , a
simple machine learning algorithm.

 The user provides with range of fluctuations of the market capital


value of the bitcoin for a given day. Based on these values further
approximate market capital of the bitcoin in predicted.

 This project has considered a large historical dataset, with an


approximate count of 5500 records for further prediction.

 This project aims at accurate prediction of the market capital value


of Bitcoin , for financial investors and other consumers.
ABOUT THE COMPANY
 Vision Digital India in association with Imperial Group for 10 years have
inspired to provide digital, analytical and app developing skills in the
present generation to meet the corporate needs. To support Digital India
initiative, Vision Digital India is offering courses like Certified Machine
Learning Course in association with Google, Certified Analytics for all in
association with IBM and iOS Mobile Application Development in
association with Apple.

 They offer first in its kind of most innovative programs on Machine Learning
and Data Analytics
INTRODUCTION
 A glimpse about Machine learning.
 Machine learning is an application of artificial intelligence (AI) that provides
systems the ability to automatically learn and improve from experience
without being explicitly programmed. Machine learning focuses on the
development of computer programs that can access data and use it learn
for themselves.

 The process of learning begins with observations or data, such as examples,


direct experience, or instruction, in order to look for patterns in data and
make better decisions in the future based on the examples that we
provide. The primary aim is to allow the computers learn automatically
without human intervention or assistance and adjust actions accordingly .
APPLICATIONS OF MACHINE LEARNING

• Virtual Personal Assistants


• Predictions while Commuting
• Videos Surveillance
• Social Media Services
• Email Spam and Malware Filtering
• Online Customer Support
• Search Engine Result Refining
• Product Recommendations
• Online Fraud Detection
AN INSIGHT ON THE PROJECT
 GOAL STATEMENT
Bitcoin is a cryptocurrency used by many countries for business transactions
and financial investments. There are few contributing factors on the bitcoin
pricing , some of them are : supply of bitcoin and market for it , competing
cryptocurrencies , trades done on bitcoin etc. Inconsideration with all the
above factors we predict the market capital value for the given volume of
Bitcoin.
BITCOIN MARKET CAPITAL ANALYSIS
 Stock market and financial investment field would never lose its
demand and importance in the present world.
 Efficient money investment is the game changer and every
individual wants to be aware of the factors affecting these
changes.
 This project is used for live prediction of the Bitcoin market capital
for each day , which is essential stock market and investors.
 The prediction is important for efficient investment and analysis.
 These insights can be drawn by applying machine learning
algorithm – Random Forest.
MACHINE LEARNING TECHNIQUES
Discussion of algorithm
 We have used random forest algorithm also called as random dicision

forest.

 One of the popular ensemble method.

 Can be used for both prediction and classification

 Build forest of unrellated decision trees.


LITERATURE SURVEY
 We have considered various papers while considering the working of the
project, various papers have given a profound understanding of the
methodology used in prediction system in Machine Learning.
 Technical paper by Smitha, SRM university has given the best
understanding of the methodology used in data preprocessing
 Another technical paper which is given by K. Hiba Sadia, Aditya Sharma,
Adarsh Paul, Sarmistha Padhi, Saurav Sanyal on the topic “Stock Market
Prediction Using Machine Learning predictions” where the main objective
of this paper is to find the best model to predict the value of the stock
market.
 Courses by Andrew NG in course era has given the basic understanding of
the classification and regression models in a better way.
REQUIREMENTS

 HARDWARE REQUIREMENTS
Processor : Pentium IV processor

Processor Speed : 2.4 GHz

RAM : 1 GB

Storage Space : 3.8 GB


SOFTWARE REQUIREMENTS

1. Operating System: Windows 10


2. Web browser
3. Anaconda Navigator
4. Jupyter Notebook
5. Python3
6. Libraries: numpy, pandas, sklearn
SYSTEM DESIGN
DETAILED DESIGN

Bitcoin market capital predictor is a virtual predictor which is developed to


have a enhanced way of predicting the market capital value for a given
volume of bitcoin, this interface is built in a interactive version which allows you
to provide certain inputs required for prediction and gives outputs based on the
values entered using certain machine learning techniques which evolved in
previous times.
DATA COLLECTION

 The data was collected from Kaggle website.

 The project initially started by collecting the necessary data


regarding the Bitcoin prices, we have collected complete report
having the information on the highest value reached on that
particular day, lowest value reached on that particular day,
opening price of the bitcoin value on a particular day, closing
price of the bitcoin value on a particular day and bitcoin volume
considered. We have considered daily records from April 2013 – July
2019 having total records of 5500 in a csv format which is
completely elaborate on all the attribute details.
Snapshot of dataset.
Data Set Attributes

 Date : Date of that day

 Open : Open price at start time window

 High : High price within time window

 Low : Low price within time window

 Close : price at end of time window

 Volume : Amount of Currency transacted in time window

 Market Cap : Market Value

 Here we use Market Cap as the dependent variable and the remaining attributes

are the independent variables.


Import Libraries
Following are important class of libreries that we need to consider for this

project
IMPLEMENTATION
 After collecting the data from dataset , we prune the dataset to remove
any noise, unwanted entries , fill the null data etc.

 This is the code to read the dataset on to the python file.


We replace the empty cells with mean of the particular attribute. It is one
of the methods to handle and standardise the dataset.
 To evaluate the given model , we used simple machine learning algorithm .
 Linear regression algorithm was implemented on to the dataset and it was
observed that the model provided very less accuracy of 29%.
 Random decision forests correct for decision trees' habit of overfitting to
their training set. The code snippet for random forest is given in the below
figure xxx and it provides the highest accuracy among other algorithms that
is 99.96%.
Methods used to improve the model
Output generated was :
CONCLUSION

With rapid growing technology, it is essential to be on par with the


requirements of the users. It is of outmost importance to accept the
increasing demands of the users and to enhance the existing
methods. The internship project help us realize this need in real world.
Understanding the cryptocurrency like bitcoin is not easy to layman,
thus helping one in need was fulfilled by this project. With reference to
this prediction one will come to understand the pattern behind the
working of bitcoin and can use it in their business purpose.
FUTURE ENHANCEMENT

Regarding the future enhancements the project can be implemented


with more pictorial representation and one can glance the concept
even better. With growing data day by day the pattern also changes
in prediction. Taking into consideration of this fact we can add
mechanism to take live data to keep the machine updated. Also we
can generate an user friendly interface for the users to input the
affecting factors of bitcoin pricing and obtain much accurate results.
QUESTION AND
ANSWER
THANK YOU

Das könnte Ihnen auch gefallen