What the AI Score means, how it is built, and the universe it covers.
Every qualifying stock gets one AI Score from 1 to 10. It is an ensemble of five sub-scores, so it reads a name from several angles at once rather than a single indicator. Higher is a stronger positive edge; the score maps to a plain verdict.
Strong positive edge across the signals.
Mixed or neutral: no clear edge either way.
Weak signals; the odds lean against it.
Each sub-score is also on a 1-10 scale. They combine into the headline AI Score, and you can see them per stock on the board.
A score reflects a stock's probability advantage of beating the DSEX over the next three months: the gap between its estimated probability of outperforming and that of an average listed share. A positive advantage lifts the score, a negative one lowers it. The board shows this advantage per stock so the number is never a black box.
Each stock lists a few alpha signals: the driver features that most influenced its score, such as price above its key moving averages, an attractive valuation versus its sector, or below-average volatility. They are the "why" behind the number.
The board also shows how much a stock's AI Score has moved over the last seven sessions. A rising score can flag improving conviction before it is obvious in price. No other DSE site publishes this.
To keep rankings meaningful, only liquid names are scored as top picks: market cap above ৳500 crore and a three-month average volume above 100k shares. Z-category stocks are excluded by rule. Illiquid names are not ranked at the top.
Signal IQ is generated by our models from price, volume, and financial data. It is educational decision support and research, not personal financial advice, and it does not guarantee any outcome. Scores and levels change daily; the decision, and the risk, are yours.
