The Masternode ecosystem continues to grow, with over 500 Masternode coins and counting as of this writing. This is up more than 150% from a year ago. Though the trend in the number of masternode coins has been steadily increasing, the market capitalization of this sector has exhibited its own unique behavior that has consisted of large swings in valuation. Naturally, investors have a vested interest in understanding past, present, and future trends in the Masternode ecosystem, which can help inform investment decisions. In this article, we've used four machine learning techniques to generate one-month forward projections of the Masternode sector’s market capitalization.

To fit these models, we began by collecting Marketcap data from masternodes.online. We then supplemented our dataset with the overall Marketcap from the chart from Masternodes.Online. From there, we cross-referenced the DASH Marketcap over this time period and removed this from the overall total as an attempt to measure the overall ecosystem without DASH's particularly large influence. This gave us data from February 21st, 2018 until the end of the year, a little over 10 months’ worth.

We then used this data to fit four models:

  1. Linear Model - This model fits a linear trend to the previous data points in order to project January 2019. This is the simplest model of the bunch.
  2. ARIMA - This model accounts for the variance across time and uses a windowed (or moving) average to predict future Marketcap values.
  3. Brownian Motion - Looking at the changes in Marketcap each day over time (returns) we randomly sample from this distribution to simulate future returns, in order to generate projections.
  4. Prophet - Here we use seasonal trends and components of the time series data to predict future results.

Using those four techniques, we generate the following projections:

chart1

chart2-1
chart3
It can be seen that all models are projecting a downward trend during January 2019, though the intensity and the confidence in these models varies. ARIMA and Brownian Motion are relatively the most bullish, with only a minor downward trend in Marketcap projected, and large error bounds. The linear model and the Prophet model see a steeper downward trend and are a bit more confident in the projections.

The error bounds are a key aspect of these projections. For example, though the Brownian Motion model has an average projected Marketcap at the end of January 2019 as about $440,000,000 (down from ~$490,000,000 at the end of December 2019), a 95% prediction interval indicates that this value could be as much as $1,000,000,000 or as little as just under $200,000,000 by the end of January. Quite the range, but reasonable considering past observations.

Overall, our models are cautiously pessimistic about the state of the Masternode ecosystem in the coming month. However, due to the large variance in projections, there are reasonable bounds suggesting that overall ecosystem market capitalization could improve. From a statistical perspective, we would like to see the actual ecosystem market capitalization end up close to our projected red lines for one or more of the models, but due to the variance in the predictions, outcomes within the error bounds will be considered successful as well. We will check back in to assess our results at the beginning of Feburary and provide a new round of machine learning based projections.

If you like this content, follow Omni Analytics on Twitter at @omnianalytics