Topic: From SpitfireABC - please critique my V ROUGH ICO Sentiment Analysis unedited (Read 79 times)

Sentiment Analysis - Rough Initial notes by S Gledhill (SpitfireABC)


wiki link: https://en.wikipedia.org/wiki/Sentiment_analysis


- Often applies to the attitude of the speaker,
the grammer, cadence, enthusiasm, knowledge level, etc. all the different nuansces of
 'good' language usage = the reaction of the reader.
 -The polarity - positive/negative reaction etc.


- Is the text on the topic reflecting and inducing on the reader, a positive or negative view?
Ultimately, regarding the ICO.


Largely investors are emotional investors - investing based on their belief, their hopes + fears, their exitement, etc. - confidence in certain crypto currencies - BTC and Ethereum have a strong base of supporters believing that they will continually improve and go up in value. At the moment there are very,very few tools/market fundamentals to assess the value such as there are with the stock market.
Just not there yet.

Market movement is laregly based on investors confidence/sentiment/perception. (Like I always said
from the off  - Business IS Trust. (I'm trying to earn it, and learn, and tbh am enjoying it)

- Positive (or negative) social network talk largely drives the perceived value, (Eg Reddit, Bitcointalk, etc)
and therefore outcome of value real and perceived - whether the ICO goes up or down in value.

Sources on Sentiment Analysis - : - white / academic papers

"Stock Trend Analysis Using News Sentiment Analysis" - Department of Computer Engineering in Mumbaii

- Another one - Stock market prediction analysis - Stanford Uni

_____________________

The way it essentially works is that people ARE predictable.(behaviorally) (my psycholgy A level 23 uears ago wasn't a total waste of time! EG, one bad/good rumour, can go viral, and
greatly effect actual value.

_____________________

coingecko lists new icos - as well as otheres, cointracker, etc.

_______________________

Social Mention  - is a very useful tool to use for sentiment analysis.

- To use type in the ICO name.
- It lists a basic Sentiment rating and results listed ie...

Strength - how often * is being talked about
Pashion - If an ICO is being repeadly being talked about
Reach - How many people are actually talking about said ICO

Overall Sentiment - Basic breakdown of ration of how positively people are talking about * ICO

There is further information, on site.

If you then get, say a list of 100 ICOs from coin gecko, run Social Mention Analystics on your spreadsheet,
then sort on dofferent indicators, ie stratify them. via, 'strength'/ 'positive mentions' etc.
and look at these metrics, and start combining these scores, to create your own 'score' metric
and stratify further,

So, say 100 ICO's down to 5 ICO's.

Thus, sorting the ones that are likely to fail or not. I'm doing an experiment using this method,
and so far, using datasets from ICOs that have actually been launched fully, from tracking websites,
- there's many.

SG.

Criticism v welcome,

- As would be getting beyond being a 'Newb' member on here ):: - can't do bounties, etc.

Regards.

Ps. How DO i become beyond a 'Newb' because, I feel that's becoming shall we say an innacurate description of my blooming understanding and enthusiasm -and that would help more.