Datathons SolutionsTeam solutions

Deciphering Cryptocurrency by Occam’s Data Ninjas


Occam’s Data Ninjas

Please find the attachment for the code in the above google drive link.

Occam’s Data Ninjas

Team Mentor: Dr. Subhabaha Pal @drsubhabahapa
Team Members:
Team Motto: Simplicity at it’s best.
Team Toolset: R Studio, Microsoft Excel.
Business Understanding and Objective: There is need to predict the price variations of crypto-currencies such as bitcoin over time to understand the trend and to compare the variation in trends of different crypto-currencies. Another objective is trading crypto-currencies predicting the variation pattern in the next time points. We use the past trend history of cryptocurrencies, analyze them to forecast the future values by building a prediction model with the best possible of accuracy.
Data Understanding: The data was provided in the format of .csv files. The data contains the information regarding the 687 cryptocurrencies, whose individual parameters such as opening value, closing value, low, high, volume are provided as individual files. There are 2 Consolidated files that have the summary information regarding the price of each bitcoin captured for every 5 minutes interval from 17th Jan 2018 to 23rd Mar 2018. The ID’s of each crypto-currenciy and the names of each crypto-currency are provided in another consolidated file.
Data Preparation: The data provided in form of the .csv file should be qualified to be a time series. But if we observe the data, it has lot of missing observations that are not captured at regular intervals of time. The main and basic requirement of Time Series is data should be captured at regular intervals of time which are equally spaced. So we need to impute the missing values to get the best prediction model.
Identifying the Missing Values: First we have created a vector having the sequence of timestamps from the starting time in data to the ending time in data with a equal spaced interval of five 5 minutes. Now we have performed the outer join of the vector with the actual data set. When we perform this operation, if there is match of timestamps in both the vector and data set then the values in data set are considered, but for the timestamps, which is present in the vector but not present in the data set, we will have an empty value for the all columns in the row. In the next section, we deal with filling these Missing Values.
Imputing the Missing Values: Now we have designed a trivial approach in filling the missing values for the empty rows. Let X(t) be the missing value then, it will be imputed by AVG(X(t-1),X(t+1)). Let us consider another scenario. Let X(t),X(t+1),X(t+2) be the sequential missing values then we will compare the values at X(t-1) and X(t+3), If there is increase in value from X(t-1) to X(t+3) then the amount of increment will be proportionally distributed among all the missing values. Suppose if there is a K amount of increment the K/5 is for first missing value , 2K/5 is for next missing value and 3K/5 is for the last missing value. The same logic is applied even in the case of Decrements.
DataModelling: We have used ARIMA Model to forecast the time series data. Auto Regression (AR) will determine the relation between the previous values and the current value in the time series. Moving Average (MA) summarizes the relation of the error term that appear in each observation. Integrated in ARIMA refers to the order of differencing that need to be performed to make the time series stationary. To build the model we have started in this way.
To Predict values of Y(t) (which is the first data-point in a day), we use all the data from Y(0) to Y(t-1) and form a time series with them, and next we find the order of AR, order of MA & d-value that best fit for the time series. We check the ACF and PACF values to determine the the parameters for the ARIMA(p,d,q) model and passed the ARIMA model to forecast the future values. After that we take Y(1) to Y(t) to predict Y(t+1) and we carry on these activities till all 288 data-points in a day are predicted. 
Evaluation: We have divided the data into  2 parts namely training and test data set. we have fed the model with the train data set that helped the model to learn something about the data. Now we have predicted the values present in the train data set. We have compared the Predicted values with the Observed values and calculated the RMSE( Root Mean Square Error). We have selected the model which has the least RMSE Value and higher Accuracy.

Share this

3 thoughts on “Deciphering Cryptocurrency by Occam’s Data Ninjas

  1. 0

    Are you looking for a profitable investment where you can start with a little amount and earn a reasonable profit within a short period of time?. I never believed in any online investment because I was scared and never wanted to be cheated, until I saw a review about Mr Pablo Martinez. He’s a Forex/Crypto trading account manager who can help you manage your trading account with his trading strategies and winning signals. I started with an investment of $500 and earned a profit of $6,650 within 7 days. I now earn quite a lot on a weekly basis and I owe everything to Mr Pablo Martinez. Thank you Mr Pablo Martinez for turning my financial life around, and I will keep recommending your good works. If you want to invest in Stock, Binary options and Forex/Crypto trading, kindly contact Mr Pablo Martinez and you’ll be glad you did. There are no hidden charges.

    Contact Mr Pablo Martinez through

    E-mail: [email protected]

    WhatsApp: +44 7520 636249


  2. 0

    Thanks to Mrs Jane who helped me recover all my lost funds in forex and crypto trading including my profits, i was a big fool giving my hard earned money to greedy and scammed brokers, but am so happy i met Jane silva a honest woman who helped me recover all my lost funds, and she also gave me the right signal and platform to trade with, now am able to make $5000 weekly, and am very happy, that is why i cant stop testifying about her, if you are out there still experiencing failure trading in binary option, crypto and forex trading or you want to recover your lost funds trading in binary/ forex trade i will advice you to reach out to
    her via email on   janesilva0727 gmail com

  3. 0

    ‘’I need back my money’’.. ‘’ I need my family back’’ that was the only thought I had for months and thanks to Amanda , I got back my family and my money .. I was depressed for months as my husband and 5year old left me to go live with his mum cos I used all our savings to invest in crypto investment company, took a loan and sold my car too in a bid to pay the withdrawal fees I was desperate very desperate and I nearly lost everything but thanks to Mrs Amanda , She recovered my $365k from those heartless scammers .. it’s a long story but at the end I was happy , I am forever grateful to her.. This has made me cross paths with her through a review here and am writing this review here in hope that it will help someone out there. If you’ve been in a similar situation please reach out to Amanda, She is very competent and reliable . The contact details are as follows , email: Amandaeverbrant01 on Gmail or  whatsapp +1 (562) 543‑3882

Leave a Reply