Discover the world's scientific knowledge

I have come across many research articles stating the values of rmse of deep learning models such as LSTM. Some results are between 0 and 1, while some are between 0 and 50. I want to know what an ideal rmse value is for such neural network models.

I have been trying to reduce the rmse values to less than 1 from values between 3 and 7.

Note: The error values are computed after unscalling back the dataset to its original form. The errors are, however, between 0 and 1 when computed with scaled data. But, I think computing with unscaled data makes more sense.

I am doing simulation in NS2, I need to get the RSSI value in Mac-802.11.cc to modify the client table access-point in NS2. Can any one help me to modify or add the RSSI value in the client table?

I am interested in calculating transmission power in case of underwater sensor networks. I am calculating the transmit power as follows. Kindly suggest if my calculation is right. Moreover also suggest whether the values i am using i.e values for bandwidth, spreading etc are correct.

Absorption and transmission loss models for underwater sensor networks are given as follows.

absorption=10^((0.11*f^2/(1+f^2))+44*f^2/(4100+f^2))+2.75*10^(-4)*f^2+0.003))/10);

TL=(d^spreading)*(absorption^d);

Where the Power in dB and power in watts is calculated as follows

AcousticPower=NoisePDF*TL*Bandwidth*TargetSNR;

TX_Power_in_watts=(AcousticPower*10^(-17.2))/circuitefficiency;

The following values are being used.

f=25;

Bandwidth=1;

spreading=1.5;

TargetSNR=15.4120;

NoisePDF=10;

PWreception=1; (kindly suggest if the value for reception power level(PWreception) is correct or not)

Best Regard

I am currently using back propagation to train a ANN model that simulates a swinging pendulum. The input of the system becomes saturated as it reaches a maximum limit value. What method should I use to train the ANN model so that the output doesn't also become saturated?

For sensing a real time signal by using SDR what is the procedure to follow to set the threshold value. In most of the documents its mentioned as trial and error method. Is it the average of the power. But then if noise power is more  or less then it leads to mis-detection  and false alarm. Can any one suggest a way to set the threshold to detect a real time signal using SDR?

I want to know how we can set the epoch value and after training what will be the best range of performance value?

I am trying to use one class SVM on a data set of a person's hand motion when touching a smartphone screen(I coded with android). I am trying to make a model that decides if the given hand motion is of the owner's or an intruder's.

I have a question about the nu parameter. I read 'nu upper bounds the fraction of errors and lower bounds the fraction of SVs, and that increasing nu, i.e. allowing more errors, increases the margin'. Does this mean that if the data set I am using has no error(outlier), which means all of the data are correct ones from the smartphone owner's, then I can put nu to a very low value?

I got 84% correctly classified when nu is 0.1, 89% when nu is 0.05 and 92% when nu is 0.03. The samples are correct ones from the smartphone user and the number of features are 10(I put gamma to 0.1). Is the nu value too low?..

I'm looking for actual value that represent the distance between the sensor node and gateway in the wireless body area network (WBAN). Can anyone help !

It is not found the well know value of VMS parameter traffic per user. Some says its 2 milli erlang some says it is 0.2 erlang. Being the adjustable parameter, though want to know the standard and empirical value.

If table is like can we draw venn diagram by this input?

variable          value r1               value r2

a                        20                         45

b                         10

c                          39                        1

d                                                      37

e                            98                     27