Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 78 views

    Project Based Learning: Predicting Bitcoin Prices Using Deep Learning

    Uploaded by

    Suraj Bahadur
    This document describes a project to predict bitcoin prices using deep learning models. The objective is to predict bitcoin prices based on real-time data. Existing methods like SVM and sentiment analysis are discussed along with their drawbacks. The proposed system uses deep neural networks and LSTM for prediction.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Project Based Learning: Predicting Bitcoin Prices Using Deep Learning

    Uploaded by

    Suraj Bahadur
    78 views24 pages
    This document describes a project to predict bitcoin prices using deep learning models. The objective is to predict bitcoin prices based on real-time data. Existing methods like SVM and sentiment analysis are discussed along with their drawbacks. The proposed system uses deep neural networks and LSTM for prediction.

    Original Description:

    ppt

    Original Title

    BATCH_30 (1)

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document describes a project to predict bitcoin prices using deep learning models. The objective is to predict bitcoin prices based on real-time data. Existing methods like SVM and sentiment analysis are discussed along with their drawbacks. The proposed system uses deep neural networks and LSTM for prediction.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    78 views24 pages

    Project Based Learning: Predicting Bitcoin Prices Using Deep Learning

    Uploaded by

    Suraj Bahadur
    This document describes a project to predict bitcoin prices using deep learning models. The objective is to predict bitcoin prices based on real-time data. Existing methods like SVM and sentiment analysis are discussed along with their drawbacks. The proposed system uses deep neural networks and LSTM for prediction.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    of 24

    SIDDHARTHA INSTIUTE OF ENGINEERING & TECHNOLOGY

    (Affiliated to JNTU, Hyderabad)


     Vinobha Nagar, Ibrahimpatnam, RR– 501 506, Telangana, INDIA.
     

    PROJECT BASED LEARNING : PREDICTING BITCOIN PRICES


    USING DEEP LEARNING
    by

    M. RAMYA SREE 16TP1A0566


    NANDINI SETTY 16TP1A0587
    SURAJ BAHADUR KUMAL 16TP1A0591
    Y. MANIKANTA 16TP1A05A5
    Under The Guidance
    Of

    Dr. R . Kavitha
    OVERVIEW
    • INTRODUCTION
    • OBJECTIVE
    • EXISTING SYSTEM
    • PROPOSED SYSTEM
    • PROPOSED MODEL
    • UML DIAGRAM
    • MODULES
    • HARDWARE REQUIREMENTS
    • SOFTWARE REQUIREMENTS
    • RESULT
    • CONCLUSION
    • FUTURE ENHANCEMENT
    INTRODUCTION

    • Project based learning is the methodology in which


    projects drive knowledge.

    • Project uses Deep Learning Methodology.


    OBJECTIVE
    • Predicts the price of a cryptocurrency
    which is bitcoin

    • We attempted to predict the prices of


    Bitcoins based on the real-time data
    and using deep learning
    methodologies.
    EXISTING SYSTEM
    • SVM
    It is a classification method.

    • Sentiment Analysis
    Predicting Bitcoin price by predicted trading volume
    using Google Trends views
    Drawbacks

    Sentiment analysis is not


    SVM results in Low considered to be a factor
    because of greater risk of
    Accuracy
    manipulation.
    PROPOSED SYSTEM
    Deep Neural Network [DNN]
    Deep Neural Network (DNN) - based models performed the best for price ups and
    downs prediction (classification).
    Neural Network with Hidden Layer
    PROPOSED MODEL
    Recurrent Neural Network

    • Can memorize previous input in memory when


    sequential data is given

    • Works on Backpropogation
    BACKPROPAGATION
    • If actual output and expected output differs ,
    the weight gets corrected itself

    Error (e) = (actual output - model output)2

    Problem with Backprogration is LONG TERM DEPENDENCY


    1. Vanishing Gradient
    2. Exploding Gradient
    Long Short-Term Memory (LSTM)

    • Designed to avoid Long Term Dependency


    Problem

    • Capable of remembering Information for


    longer periof of time
    Use case
    Diagram
    MODULES

     Data Collection
     Data Preprocessing
     Convolutional Neural Network
     Recurrent Neural Network — LSTM
    Advantages

    Easy to Implement with better DNN-based prediction models performs


    accuracy the best.
    Hardware Requirements

    PROCESSOR: intel i5 RAM : 4GB RAM HARD DISK : 250 GB

    Software Requirements

    OPERATING IDE : SPYDER3


    LANGUAGE: PYTHON
    SYSTEM :WINDOWS
    10
    A four step process to Predict Bitcoin prices

    • Getting real-time crypto currency data.


    • Prepare data for training and testing.
    • Predict the price of crypto currency using LSTM neural network.
    • Visualize the prediction results
    Step 1 : Getting real-time cryptocurrency data from online website using API
    calls
    This is how the dataset looks like with all the trading features like price, volume, ask, high, low.
    Step 2 : Prepare data for training and testing. 

    Splitting the data into two sets — training set and test set with 67% and 33% data respectively.
    Step 3 : Predict the price of cryptocurrency using LSTM neural
    network
    Step 4 : Visualize the prediction results
    Conclusion
    • Predicting the future will always be on the top of the list of uses for
    machine learning algorithms. Here in this project we have attempted
    to predict the prices of Bitcoins using two deep learning
    methodologies.
    Future Enhancement
    • Train the model on a larger data set to increase prediction accuracy.
    • Design model with high number of neurons and run on a
    supercomputer or a cluster of system.
    THANK
    YOU

    You might also like