+- kanji koohii FORUM (http://forum.koohii.com)
+-- Forum: Learning Japanese (http://forum.koohii.com/forum-4.html)
+--- Forum: Learning resources (http://forum.koohii.com/forum-9.html)
+--- Thread: @cangy: Qs about your kore chart (/thread-5438.html)
@cangy: Qs about your kore chart - gfb345 - 2010-04-17
cangy, add me to those who are thankful for your kore resource!
I'm trying to understand this chart from your page.
What do "nuke" and "fancynuke" mean?
I'm particularly intrigued by the fact that the curves for "unsorted", "nuke" and "fancynuke" all come together at two points, but do not cross, which is somewhat unusual.
Also, I'm puzzled by the long constant stretches in the "fancynuke" curve. (There's also a much less striking constant stretch in the ko2001 curve.)
Finally, what's up with "const"? That's not really a sort field, is it?
TIA!
Edit: fixed typo (cont => const)
@cangy: Qs about your kore chart - Kubelek - 2010-04-17
const = constant
@cangy: Qs about your kore chart - gfb345 - 2010-04-17
Kubelek Wrote:const = constant"constant" what?
@cangy: Qs about your kore chart - Kubelek - 2010-04-17
it just represents a linear growth, I think it's there for illustration purposes. I assumed it meant a constant function, which sounded intuitive, although I have no clue if that's what it's called in English. Sorry.
@cangy: Qs about your kore chart - Fuamnach - 2010-04-17
the number of unique kanji is constant. on the graph it looks like for every 1000 sentences there are approx 280 new kanji and this doesn't change, so at 2000 sentences there will be 560 kanji and so on.
I think...
Edit: gah, Kubelek beat me...:)
@cangy: Qs about your kore chart - Kubelek - 2010-04-17
after I reread his original post I started doubting my explanation and I changed it, so it's cool
@cangy: Qs about your kore chart - Fuamnach - 2010-04-17
LOL, I know that feeling
Another thing, I think the long constant stretches you refer to in the "fancynuke" (which, btw, I have no idea what this actually is :/) mean that at around 400 sentences (corresponding to the start of core 2k's hideous step 3?) there are no new kanji till the 1000th sentence, then a whole bunch you have to learn to till you finish core 2k. Then for the next long stretch no new kanji till almost the 5000th sentence then you have to learn a whole bunch again till you finish. The ko2001 line just means it's the closest out of all of them to a constant progression of new kanji.
I could be completely wrong though. Best wait for cangy to tell you the truth
@cangy: Qs about your kore chart - cangy - 2010-04-18
the slope of the graph is the rate of new kanji introduction per sentence, so a flat horizontal line is no new kanji being introduced
nuke is nukemarine's suggestion of dividing into 3 sections, core2k steps 1-2, core2k steps 3-, and core6k, and sorting by ko2001 order within those sections
because unsorted and nuke cover the same sentences within those sections, they have to meet at those points
from the discontinuities in the unsorted slope at those points, you can see that they are the only real divisions within the original ordering
fancynuke is the same division into 3 sections, but within those sorting by kanji used in the previous section (sorted by frequency in the current section), then the remaining kanji in the current section by frequency, basically delaying introducing new kanji as much as possible and hence flattening the graph
I assume the flat section in ko2001 is due to finally reaching a common kanji that is used in many sentences containing only previously seen kanji
const isn't an actual sort order, but is the average rate for the whole collection, and is there as a reference for what I think would be the ideal -- a constant rate of new kanji introduction
@cangy: Qs about your kore chart - Nukemarine - 2010-04-18
**Edit: guess I replied while Cangy was posting his. Eh, it happens**
Well, I think he's referring to my moniker since I was asking about sorting Core 2k/6k according to 2001KO order.
The "Nuke" is sorting the entire Core 2k/6k according to 2001KO order (then RTK if into 2k1KO). The "Fancy Nuke" refers to the list where you study the first 400 sentences of Core 2k (sorted), then the next 1600 of Core 2k (sorted), then the Core 6k sorted. It works on the idea that there are common words using more advanced Kanji that should be learned before the less common words using simpler kanji. Of course, that means you don't get new kanji in later groups of sentences till a bit later in Core 2k steps 3-10 and even later in Core 6k.
@cangy: Qs about your kore chart - gfb345 - 2010-04-18
@cangy, @Nukemarine
Many thanks! That clarifies a lot!
@cangy: Qs about your kore chart - cangy - 2010-04-22
Nukemarine Wrote:The "Nuke" is sorting the entire Core 2k/6k according to 2001KO orderactually, that's ko2001...
Nukemarine Wrote:The "Fancy Nuke" refers to the list where you study the first 400 sentences of Core 2k (sorted), then the next 1600 of Core 2k (sorted), then the Core 6k sorted....and that one is nuke