Commit Graph

55 Commits

Author SHA1 Message Date
a038c2f696 markov models and iei interval analysis completed 2023-06-19 15:28:53 +02:00
3a4ece9902 savety 2023-06-19 14:33:32 +02:00
988d80f7bd work in progress 2023-06-16 15:37:54 +02:00
d11c83b46f savety commit 2023-06-16 11:14:59 +02:00
cabb3bd2c9 savety commit 2023-06-16 11:12:35 +02:00
e22108cdd7 add void to ethogram. ethogram is now colorful :D 2023-06-16 10:52:09 +02:00
725c0b426a make chirp-chrip cross correlation. marcov model works but requires refinement. arrows from each node should sum up to 100pct ?! 2023-06-14 15:05:49 +02:00
8af8c92b27 abs dEODf correlation with lose chrips and med chase duration. chirps look like mexwell distribution. 2023-06-14 10:09:20 +02:00
96cb4b607c dada 2023-06-13 14:28:04 +02:00
e2c284ea76 adapted code in a way that rise analysis include more recordings. chirp analysis skip those marked as bad by patrick 2023-06-13 09:22:33 +02:00
50c762ab8d adapted code to compute statistics correctly: trial_mask. done for trial_summary_eval.py. The rest is missing. DO IT. Next and last script: marcov model 2023-06-12 15:28:32 +02:00
3122cb10dc trial summary now includes information about agonistic counts and competition duratin ala ana silva. started to adapt trial_summary_eval.py accordingly to evaluate this 2023-06-09 13:23:43 +02:00
47a1ac1058 forgot the commit and push on wednesday 2023-06-09 07:58:56 +02:00
07daf7457c alot of statistics 2023-06-06 16:56:16 +02:00
bb45677bd1 analysis almost complete 2023-06-05 15:40:46 +02:00
11b58dda36 event vs. time pir charts done 2023-06-02 11:04:08 +02:00
14c7538cb9 event time analysis continued... not much left 2023-06-01 14:53:05 +02:00
f096e9ba4f inter event interval analysis done. make function more universial. relative rate progresion of ecent evaluated 2023-06-01 10:46:13 +02:00
4c013ab869 time series analysis... excluded chirps in event time correlations. maybe check plots again 2023-05-31 10:49:29 +02:00
aa200cf049 changed name of time correlation script to event_time_correlations.py 2023-05-31 09:25:38 +02:00
32bcb3aeea temporal correlations for all logical combinations?! 2023-05-31 08:06:16 +02:00
e3a5cadc0c works nice... adapt loop and clean up the code. 2023-05-30 15:03:48 +02:00
fbad4ffa4b plot looks nice. 2023-05-30 13:15:45 +02:00
4d045d6ed0 cross correlations now work... alter code in a way that time axis can be defined out of functions. inlude raster. adapt code for other event combinations 2023-05-30 11:45:02 +02:00
90424c2c46 added meta files for convinience 2023-05-30 07:07:29 +02:00
7a5c5cfd0e bf. Monday I need to look into the norming of the kde curves... ToDo comment in one section 2023-05-26 15:34:41 +02:00
aec5e298bc plot 2023-05-26 15:33:23 +02:00
8a9ec1beed bf 2023-05-26 15:20:12 +02:00
cb37108220 bf 2023-05-26 15:17:58 +02:00
caa01a3005 bf 2023-05-26 10:42:25 +02:00
0e36d12419 bug fix 2023-05-26 10:33:16 +02:00
d6448209e0 bug fix 2023-05-26 10:32:06 +02:00
b80a2c8de9 print statement to assess bootstrap speed 2023-05-26 10:30:35 +02:00
11c015b1c2 bug fix 2023-05-26 10:26:29 +02:00
9cc35819a0 bug fix 2023-05-26 10:24:48 +02:00
399197d2b1 bug fix 2023-05-26 10:09:26 +02:00
6d7e58ef80 boot implementations using cupy 2023-05-26 10:06:02 +02:00
d8d28df2ea test 2023-05-26 08:40:14 +02:00
4f64aa6668 event time ananlysis started. tried to implement bootstrap with cupy ... memory issues ... resolve on monday 2023-05-25 15:39:04 +02:00
d9d9982635 meta analysis continues... next step would be to load the time series data and calculate cross correlations 2023-05-23 14:53:46 +02:00
8c1f85f89e cleaned code 2023-05-23 11:01:59 +02:00
5af9a25dfd continued meta analysis 2023-05-22 15:15:16 +02:00
42640c17d9 evaluation of the trial_summary initiated. Now the fun beginns 2023-05-19 13:17:26 +02:00
3c9bbb6a91 bf 2023-05-19 08:17:31 +02:00
83d5e3164b include mean shelter power (which denfines winner) to plot. 2023-05-19 08:15:47 +02:00
922c938fb6 save ids in trial analysis 2023-05-19 07:57:34 +02:00
38b4b6800c sex assignment works. maybe rewrite trial analyis in a way that everything is saved with reference to ids not winner/loser... seems to be fishy. 2023-05-17 15:15:15 +02:00
c7725c4e01 meta data collection nearly finished. Need to assign sexes based on EODf25 mean of each fish at the end. I got issues with winner detection in the trial_analyis.py --> results inconcurrent with csv values. plot mean/median shelter power used for winner assignment 2023-05-17 14:32:34 +02:00
8517014571 created DataFrame containing trail data for corrections. Next would be to get to know sexes: load baseline -> q10 comp freqs -> > 749 == f 2023-05-16 15:12:24 +02:00
fdd5e1cc01 continued on analysis ... included the calculus of mean_l and w and detected all LED_frames 2023-05-15 15:20:07 +02:00