Researching exchange-listed products using sentiment
A metric, such as price, is estimated for an exchange-listed product based on public sentiment regarding the exchange-listed product. A system for estimating a metric for an exchange-listed product receives data from an external server. The data may be received from a social media website, and may represent user-generated data in natural language form. The system identifies a portion of the data and determines whether the portion represents a positive sentiment or a negative sentiment. The system estimates a metric for the exchange-listed product based on the sentiment represented by the data. The system may combine the metric based on sentiment with other metrics to create an averaged metric. The other metrics may be obtained using analyst consensus estimates as well as technical analyses of market information for the exchange-listed product.
1. A computer implemented method, comprising:
estimating a first metric based on analyst consensus estimates, wherein the analyst consensus estimates include a first analyst consensus estimate of a price for an exchange-listed product for a first time period and a second analyst consensus estimate of the price for the exchange-listed product for a second period of time;
estimating a second metric based on technical market data;
estimating a third metric based on public sentiment analyses; and
producing a combined estimate based upon the first metric, the second metric and the third metric;
wherein the combined estimate is a price for an exchanged listed product, and
wherein each estimating operation and the producing operation are executed by a computer.
2. The computer implemented method of claim 1 wherein producing includes producing a combined estimate based upon weighted values of the first metric, the second metric and the third metric.
3. The computer implemented method of claim 1 further comprising:
providing a rating proportional to the difference between a current price and a target price of an exchange-listed product.
4. The computer implemented method of claim 3 wherein the rating is expressed as a function of:
E 1 (Forecast 1 - MP )+ E 2 (Forecast 2 - MP )+ E 3 (Forecast 3 - MP )
where Forecast 1 and E l denote a first forecast based on analyst consensus estimates and an average margin of error for the first forecast, Forecast 2 and E 2 denote a second forecast based on sentiments and an average margin of error for the second forecast, and Forecast 3 and E 3 denote a third forecast based on predictions by users and an average margin of error for the third forecast, and MP denotes the market price for the particular stock.
5. The computer implemented method of claim 4 wherein the rating is normalized to a value between −100 and +100.
6. A computer implemented method, comprising:
estimating a first metric based on analyst consensus estimates;
estimating a second metric based on technical market data, wherein the technical market data includes an actual price for an exchange-listed product;
estimating a third metric based on public sentiment analyses; and
producing a combined estimate based upon the first metric, the second metric and the third metric;
wherein the combined estimate is a price for the exchanged listed product, and
wherein each estimating operation and the producing operation are executed by a computer.
7. The computer implemented method of claim 6 wherein the technical market data further includes a second actual price for another exchange-listed product and a correlation of the actual price and the second actual price.
8. A computer implemented method, comprising:
estimating a first metric based on analyst consensus estimates;
estimating a second metric based on technical market data;
estimating a third metric based on public sentiment analyses, wherein the public sentiment analyses are derived from the evaluation of data feeds; and
producing a combined estimate based upon the first metric, the second metric and the third metric;
wherein the combined estimate is a price for the exchanged listed product, and
wherein each estimating operation and the producing operation are executed by a computer.
9. A computer implemented method, comprising:
estimating a first metric based on analyst consensus estimates;
estimating a second metric based on technical market data;
estimating a third metric based on public sentiment analyses, wherein the public sentiment analyses include positive and negative sentiment indicators; and
producing a combined estimate based upon the first metric, the second metric and the third metric;
wherein the combined estimate is a price for the exchanged listed product, and
wherein each estimating operation and the producing operation are executed by a computer.
10. The computer implemented method of claim 9 wherein the public sentiment analyses include neutral sentiment indicators indicative of public brand awareness.
11. The computer implemented method of claim 9 wherein the public sentiment analyses are derived from seed words and synonyms for the seed words.
12. The computer implemented method of claim 9 wherein the public sentiment analyses include evaluations of negation words.
13. The computer implemented method of claim 9 wherein the public sentiment analyses include an evaluation of an overall sentiment of an input string.
14. The computer implemented method of claim 9 wherein the public sentiment analyses include an evaluation of the geometric mean of sentiment probabilities of all words in an input string.