Some question in "cross_vali_data_convert_merge"

ermongroup / Wifi_Activity_Recognition

Code for IEEE Communication Magazine (A Survey on Behaviour Recognition Using WiFi Channle State Information)

GNU General Public License v3.0

251 stars 78 forks source link

Some question in "cross_vali_data_convert_merge" #30

Open beyond96 opened 3 years ago

beyond96 commented 3 years ago

What are the purposes of setting the following parameters?

window_size = 1000 threshold = 60 slide_size = 200

Many thanks!

beyond96 commented 3 years ago

"y = np.zeros(((len(tmp2) + 1 - 2 * window_size)//slide_size+1,8))" Line 46

I feel confused, Could you help me?

Hirokazu-Narui commented 3 years ago

"window_size" means the number of sampling point for the input data. (30 channels x 3 antenna x 1000 sampling points will be the input.)

"slide_size" means how many points will be overlapped for the input data. In this case, 1 to 1000, 200 to 1200, 400 to 1400... will be input data.

"threshould" means how much percent of data should be included to activity data. If 60 percent of "bed" label is included in the window, it will be "bed" label.

beyond96 commented 3 years ago

Thanks for your reply! Another question,:Would you mind shaing your code of Figure4 in your paper?Why can't I get the result of Figure 4 when I use STFT?

I am so sorry to disturb you.