Dataset Name: danel_smartscore_usa
Group: sentiment
Vendor: Danel Capital
Data Starts at: 2017-01-04 00:00:00
Symbol Set: US Equities
Asset Class: Equity,Futures
Data Update Time(s): 7:05 AM EST
Data Update Frequency: day
CloudQuant long-short backtest results show the following:
| Sharpe Ratio | Total Return | Alpha Return | Beta Return |
|---|---|---|---|
| 1.16 | 33.64 | 8.99 | 8.22 |
CloudQuant Finds Danel Alternative Data Set Provides Significant Alpha
Alpha in The Data Set
The returns for Experiment #6 : Long the top 20%, Short the bottom 4%, Hold for 20 days.
CloudQuant found that going long the top 20% and short the bottom 4% of Danel Smart US AI Signals (Danel AI Signals) returns an average of 11.20% per year after transaction costs for the most recent three years, with an overall Sharpe Ratio of 1.16 and an impressive Sharpe of 2.42 in 2020.
We found that over 40.0% of the total return is pure alpha not explained by traditional market, size momentum and value factors. The results are significant to the 89th percent level. The realized signal returns show some correlations to momentum factor. In combination with other factors, the Danel signal would be beneficial to the construction of most fundamental hedged and long-short portfolios.
White Paper Research Conclusions
We demonstrate that Danel Machine Learning Signal (Danel Signal) produces statistically significant long and short alpha at 89% confidence over a broad variety of long only and long-short portfolio structures.
The long top 20% and short bottom 4% 20-days-holding strategy returned an average of 11.21% per annum, with overall Sharpe Ratio of 1.16, a CAPM beta of 0.08, and a yearly Share Ratio of 2.42 in 2020.
The long top 4% and short bottom 4% 20-days-holding strategy an average of 11.67% per annum, with an overall Sharpe Ratio of 1.09 and a CAPM beta of 0.09, and a yearly Share Ratio of 2.68 in 2020.
The higher the Danel signal is, the stronger the alpha generated will be, resulting in the top performance in the top 20% and bottom 4% signals. The Danel Signal has a stronger long side signal than the short side.
The best holding period for Danel Signal is around 20 days.
Return decomposition analysis and correlation analysis comparing the Danel Signal to known "smart-beta" risk factors and commodities shows that over 40% strategy return is comprised of pure alpha (idiosyncratic). A more detailed model shows the alpha to be of around 10%.
The Danel Signal yields portfolios having strong positive factor exposure to momentum and negative exposure to value and consumer.
Click here to "Schedule a Demo" or Email Sales@cloudquant.com to obtain the in depth research which includes the following:
- CloudQuant White Paper - Evaluation of Danel's AI Dataset
- A demonstration of the CloudQuant algorithms
- Source Code and free access to CloudQuant Mariner to replicate the study results and/or evolve the strategies to your requirements
Data Contained in this Dataset
| Column | Type | Description |
|---|---|---|
| _seq | uint | Internal sequence number used to keep data rows in order |
| timestamp | string | Timestamp of the Data - America/New York Time. |
| muts | uint64 | Microseconds Unix Timestamp. An integer representation of a timestamp with microsecond precision that can be compared directly to other timestamps. |
| symbol | string | Trading Symbol or Ticker |
| Date | string | Date referring to the last trading day which the scores are calculate off of |
| Company name | string | Company Name |
| ISIN | string | International Securities Identification Number. An international code which identifies a securities issue |
| AI Smart Score | double | Global AI Score based on 900 data points. Range from 1 to 10, the higher the score the higher the probability to outperform the market in the next 30 to 90 days. Published at least 2 hours before US market opens. |
| AI Technicals Score | double | AI Score based on 600 Technical data points. Range from 1 to 10, the higher the score the higher the probability to outperform the market in the next 30 to 90 days. Published at least 2 hours before US market opens. |
| AI Sentiment Score | double | AI Score based on 150 Sentiment data points. Range from 1 to 10, the higher the score the higher the probability to outperform the market in the next 30 to 90 days. Published at least 2 hours before US market opens. |
| AI Fundamentals Score | double | AI Score based on 150 Fundamental data points. Range from 1 to 10, the higher the score the higher the probability to outperform the market in the next 30 to 90 days. Published at least 2 hours before US market opens. |