Using N-grams increases the dimensionality way too much.
I've thought about word-vectors as an option, but am unsure how similar terms could be grouped internally in LDA.
This kind of subject is probably worth a PHD thesis, honestly.
701 karma · joined March 26, 2015
Using N-grams increases the dimensionality way too much.
I've thought about word-vectors as an option, but am unsure how similar terms could be grouped internally in LDA.
This kind of subject is probably worth a PHD thesis, honestly.
Also, the abstraction allows for optimizations like delegating to Apache Spark or other processes/threads.
Also, there are obvious things that make developers more able to deliver quality software, and there are obvious things that make gamers more able to win games, just as there are with famous athletes. Thinking that there aren't specific predictors of success here is as silly as saying there aren't specific predictors for what makes a good QB.
The time it cost them to be software engineers is time they wouldn't have spent becoming better at sales/marketing, so your company would be in a disadvantage in the market. Even genius engineers can't overcome bad sales or marketing.
> There's not some magic thing that a senior developer has that a junior developer is missing.
"They're just better" is a way of saying "this is magic and not quantifiable". If you cannot quantify what factors allow one developer to consistently deliver quality software on schedule, then you cannot grow your team in that direction, and are essentially rolling the dice.