Need Help

Started by October 14, 2017
Can anybody provide me the solution for following problem. It will be great help.




Adaptive filter with variable update equation: Develop a stochastic gradient
adap-tive filter that attempts to minimize the following cost function:
J(n)=E{|e^2(n)|}.         e(n)<1
    =E{|e^3(n)|}.         e(n)>=1

Discuss the possible advantages and disadvantages of your algorithm over the
LMSadaptive &#64257;
On Friday, October 13, 2017 at 5:34:30 PM UTC-7, ankita bombatkar wrote:
> Can anybody provide me the solution for following problem. > It will be great help.
> Adaptive filter with variable update equation: Develop a > stochastic gradient adap-tive &#64257;lter that attempts to minimize > the following cost function: > J(n)=E{|e^2(n)|}. e(n)<1 > =E{|e^3(n)|}. e(n)>=1
> Discuss the possible advantages and disadvantages of > your algorithm over the LMSadaptive &#64257;
Discuss your thoughts so far, and I am sure you will get some help here. Do you have a partial solution, but got stuck along the way?
On Friday, October 13, 2017 at 7:34:30 PM UTC-5, ankita bombatkar wrote:
> Can anybody provide me the solution for following problem. It will be great help. > > > > > Adaptive filter with variable update equation: Develop a stochastic gradient
adap-tive &#64257;lter that attempts to minimize the following cost function:
> J(n)=E{|e^2(n)|}. e(n)<1 > =E{|e^3(n)|}. e(n)>=1 > > Discuss the possible advantages and disadvantages of your algorithm over the
LMSadaptive &#64257; Ankita, Sounds like a homework problem. You have a cost function you want to minimize. Think about what process you can use to tell if you have a minimum or not (hint: there's a reason that E[e^2] is widely used). Expand the cost function, apply the process to the expansion, and then see if that doesn't lead you to the answer.
On Friday, October 13, 2017 at 7:34:30 PM UTC-5, ankita bombatkar wrote:
> Can anybody provide me the solution for following problem. It will be gre=
at help.
>=20 >=20 >=20 >=20 > Adaptive filter with variable update equation: Develop a stochastic gradi=
ent adap-tive =EF=AC=81lter that attempts to minimize the following cost fu= nction:
> J(n)=3DE{|e^2(n)|}. e(n)<1 > =3DE{|e^3(n)|}. e(n)>=3D1 >=20 > Discuss the possible advantages and disadvantages of your algorithm over =
the LMSadaptive =EF=AC=81 I forgot to add, what does the question mean by variable update equation? (= hint: can the LMS update gain be written as a matrix?). There are numerous = examples you can research to give you ideas. After you do this research, th= ink hard about the advantages and especially the disadvantages of these var= iable gain ideas. I hope you can come up with your own idea that mitigates = the disadvantages. If you come up with something, share it with us. Maurice Givens