The last state I left the data collection and mungling. We decided to…

… use Tham's data because it's the only one with labels. However, the rest of the datasets I have a still good for the GAN.
rjm11010 · Dec 15, 2017 · 5106626 · 5106626
1 parent 65f7e1f
commit 5106626
Show file tree

Hide file tree

Showing 12 changed files with 771 additions and 1,031 deletions.
diff --git a/mkdataset/datasets/gan_datasets/.notes.md.swp b/mkdataset/datasets/gan_datasets/.notes.md.swp
diff --git a/mkdataset/datasets/gan_datasets/new.csv b/mkdataset/datasets/gan_datasets/new.csv
diff --git a/mkdataset/datasets/gan_datasets/notes.md b/mkdataset/datasets/gan_datasets/notes.md
@@ -0,0 +1,40 @@
+
+Improved motivations for group setting importance:
+
+You need to convince them to this important. You need to show that it exists in practice, or else no one is going to care about it. You need to make them care. Even if the application seems liek it's not so hot, you should describe why it is.
+You can use flocking
+
+New section for interference:
+	Describe the experimental setup
+	The goal is to see how interfence influences the experiments. You need to evaluate the effect of interference in the experiment, and it's impact.
+	So you need to display some experiement results.
+	Describe the entire experiment setup. No packages are dropped. Include how the format of the experimental.
+	You need to describe how there is no packet loss. You can see this in one of two ways. You can look at the number of packages that should have been sent, by a node, and how much were actually received.
+	You send a start signal differently depending on the scenario (aligned, staggered). Describe how you were able to synchonize their 
+	RTS/CTS are disabled, and retransmissions (retry are set to 1)
+
+Tell them the slot length.
+
+New experiment to show them the latency and loss rates based on the number of senders in a deterministic way.
+n = number of nodes
+
+1 sender and 1 receiver
+One folowing a schedule, another just listening. Tell them the latency and the loss rate of the packes.
+every 100 milliseconds send a beacon. Make a scheudle such that this happens. You should do 600 slots.
+
+2 senders, listener
+They'll be sedning simultaneously,
+Node 1 will be the benchmark for latency and loss.
+Node 2 will also be logging it's messages from 1
+
+
+3 senders
+
+5 senders
+
+7 senders
+
+
+Plot
+CDF of ONE-WAY delay for latency and loss_rate
+EAch curve should be the numberof senders (multi CDF)
diff --git a/mkdataset/datasets/gan_datasets/tham_human_and_mouse_dataset.csv b/mkdataset/datasets/gan_datasets/tham_human_and_mouse_dataset.csv
diff --git a/mkdataset/datasets/human_and_mouse_dataset.csv b/mkdataset/datasets/human_and_mouse_dataset.csv
diff --git a/mkdataset/datasets/human_and_mouse_dataset_log2_ratio.csv b/mkdataset/datasets/human_and_mouse_dataset_log2_ratio.csv
diff --git a/mkdataset/datasets/human_dataset.csv b/mkdataset/datasets/human_dataset.csv
diff --git a/mkdataset/datasets/human_dataset_log2_ratio.csv b/mkdataset/datasets/human_dataset_log2_ratio.csv
diff --git a/mkdataset/datasets/human_mouse_dataset.csv b/mkdataset/datasets/human_mouse_dataset.csv
diff --git a/mkdataset/datasets/mouse_dataset.csv b/mkdataset/datasets/mouse_dataset.csv
diff --git a/mkdataset/datasets/mouse_dataset_log2_ratio.csv b/mkdataset/datasets/mouse_dataset_log2_ratio.csv
diff --git a/mkdataset/datasets/tham_human_and_mouse_dataset.csv b/mkdataset/datasets/tham_human_and_mouse_dataset.csv