The MarketScreener IT team has updated its search system.
According to IT Team (Geek) member Magnus (first name changed), the system is "more intuitive, less cluttered, with a strong ergonomic correction on iOS devices that tended to zoom on the input.". Magnus adds, "We now use an EdgeNGram 2/25 Tokenizer coupled with two LowerCase & ASCIIFolding filters for indexing. For searching, we use HMax10 on these fields: Mnemo (*x2), name, tag, isin x0.5, text x0.1 wkn x0.1."
We didn't get it either. Fortunately, the power of AI came to our rescue. We asked a conversational AI to explain Magnus's gibberish to a 7-year-old child. Here's the answer:
"Imagine you have a magic box that can find hidden toys using special words. For the magic box to work properly, it has to learn to recognize the words, even if they're a bit jumbled or spelled differently.
The "Tokenizer EdgeNGram 2/25" is like a cutting game that takes a word and cuts it into small pieces. For example, if you have the word "balloon", it cuts it into "ba","bal" "ball", "ballo", "balloo", and "balloon". This helps the Magic Box find the toy even if you only remember a small part of the word.
The "LowerCase & ASCIIFolding" filters are like rules that say it doesn't matter if you write the words in upper or lower case, and that even if you use special or different letters, the magic box will know it's the same word.
To find the toys, the Magic Box uses a special game called "HMax10". It's like a detective who knows which clues are the most important to find the toy. "Mnemo" is like the toy's secret name, so if you know it, the Magic Box will search for it twice as hard. Names, tags and "isin" are like clues on the toy's packaging. If you know the "text" or "wkn", which are like little words written on the box, the magic box will use them too, but it knows they're not as important as the secret name in finding the toy.
So, basically, the magic box is very clever and uses all these tricks to help you find the toy you're looking for, even if you can't remember exactly what it's called."
Well, thank you AI!