Sentiment Investing Terms

We have two main terms used for trading based on sentiment data. One is the average hype index, the other is our relative hype index.

Each metric is based on the equity's previous values. Each day we scrape thousands of social media posts and assign a value between 0 and 1 to each tweet or comment. We then sum those by day, week, month, hour, etc. We use these values to calculate our metrics.

To calculate the average mentions per day, we take all the data for a single 24 hour period and sum the sentiment score (which is again between 0 and 1 inclusive). We currently use a binary classifier so all sentiment scores are either 0 or 1. Once we have the sum for the day, we divide it by the number of rows we have for that particular ticker for that particular day:

Example:

1000 total comments in a 24 hour period
500 positive (sentiment = 1) 
500 negative / neutral (sentiment = 0)
sum = (500 * 1) + (500 * 0) = 500 
average mentions per day value would then be 500 / 1000 = 0.5
Example calculation of average mentions per day

We do the above for the day, week, and month.

AHI - Average Hype Index

The average hype index is calculated as follows:

AHI = average mentions per day / average mentions per month

RHI - Relative Hype Index

The relative hype index is calculated as follows:

RHI = average mentions per day  / average mentions per week