-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathTODO.txt
20 lines (12 loc) · 887 Bytes
/
TODO.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1. Plots / grafieken verfraaien
2. Uitzoeken hoe je trendomslagen, minima, maxima kunt vinden
>> Het Quantmod package bevat de functies findPeaks / findValleys, daar lukt het misschien wel mee.
>> Het automatisch (programmatisch / algoritmisch) identificeren van pieken blijkt best een interessant wetenschappelijk probleem te zijn. Computational Topology, Persistent Homology, etc.
>> Relevante links:
https://stackoverflow.com/questions/22583391/peak-signal-detection-in-realtime-timeseries-data [*]
https://www.sthu.org/blog/13-perstopology-peakdetection/index.html
https://bookstore.ams.org/mbk-69
https://en.wikipedia.org/wiki/Persistent_homology
[*] Dit algoritme gaan we proberen toe te passen op de data; kijken hoe dat uitpakt. Zie PEAKS.R.
3. Kan de data worden gecomprimeerd? Smoothing functie? Loess???
>> Loess is hier prima voor geschikt. Zie SMOOTH.R voor details.