This is GOLD, PURE GOLD ! I wish I could have credited this course when you offered this sir. Always come back to these lectures by you. My pranaam _/\_
Thank you. The full course on "Parameter and State Estimation" that I teach currently has lot more details and broader perspectives. Of course, Kalman filter lectures makes the course more exciting.
Thank you for making such "dry" topics interesting! @14:24, how does the minimization of "sum of square of the differences" turn out to be the mean? Intuitively it seems fine, but I'm not able to derive it!
Firstly, I am happy to note that you found the lectures interesting. To your question, simply differentiate the sum squares with respect to the unknown and set it to zero. You will get the answer.
Could not understand "How the professor says the sample median is Non linear" at 21:42 . The example given around multiplication of sorted matrix is not convincing per se. Can someone help with elaboration.
He didnt mention multiplication, he mentioned a practical example of superposition theorem. The theorem states that, suppose x(t) = y. Then, x1(t) = y1 and x2(t) = y2. Theorem states that x1+x2 = y1 + y2. If you consider x as a column vector, and x(t) as a sorting function, adding them will not satisfy the theorem.
Misspecification of model means the relationship between the knowns (in this example, y[k]) and the unknowns (the constant 'c' in this example) is not specified correctly. The error in the specification could be due to a wrong assumption on the uncertainty (noise) and / or in the functional part (modelling error). As an example, suppose the generating process is y[k] = c^2 + e[k] and we assume the model to be y[k] = c + e[k]. This is the case of modelling error.
@@appliedtime-seriesanalysis7076 Thank you sir your valuable reply. This lactures are really nice and help me to understand the the concept of Estimation Theory. The way you teach us is awesome and unique. I really love your teaching. Once again Thank you very much sir.
Wow, this is pure gold.... Following from kenya
You helped me score 26/30 in my cat exam after having hated this course
Thank you for your kind words. I am glad you find this to be golden :-)
This is GOLD, PURE GOLD ! I wish I could have credited this course when you offered this sir. Always come back to these lectures by you. My pranaam _/\_
Thank you. The full course on "Parameter and State Estimation" that I teach currently has lot more details and broader perspectives. Of course, Kalman filter lectures makes the course more exciting.
@@vidyavahini1247 Dear Sir. I am unable to find the lectures. Can you please share the link?
I found my estimation theory mine,,! Started digging..!
Thank you for making such "dry" topics interesting!
@14:24, how does the minimization of "sum of square of the differences" turn out to be the mean?
Intuitively it seems fine, but I'm not able to derive it!
Firstly, I am happy to note that you found the lectures interesting. To your question, simply differentiate the sum squares with respect to the unknown and set it to zero. You will get the answer.
I LOVE YOU SIR !!!!! :) :)
Thank you for your affection!
Could not understand "How the professor says the sample median is Non linear" at 21:42 . The example given around multiplication of sorted matrix is not convincing per se. Can someone help with elaboration.
He didnt mention multiplication, he mentioned a practical example of superposition theorem.
The theorem states that, suppose x(t) = y. Then, x1(t) = y1 and x2(t) = y2. Theorem states that x1+x2 = y1 + y2.
If you consider x as a column vector, and x(t) as a sorting function, adding them will not satisfy the theorem.
Sir, could you once explain wh6b is this linear please .. 16:44
at 10:07, is c the best prediction because mean of the noise is 0? if not then what is the reason?
Yes, you are right.
Sir do you teach them in live classes?
sir, can you explain in this what is meant by misspecification of model ?
Misspecification of model means the relationship between the knowns (in this example, y[k]) and the unknowns (the constant 'c' in this example) is not specified correctly. The error in the specification could be due to a wrong assumption on the uncertainty (noise) and / or in the functional part (modelling error). As an example, suppose the generating process is y[k] = c^2 + e[k] and we assume the model to be y[k] = c + e[k]. This is the case of modelling error.
@@appliedtime-seriesanalysis7076 Thank you sir your valuable reply. This lactures are really nice and help me to understand the the concept of Estimation Theory. The way you teach us is awesome and unique. I really love your teaching. Once again Thank you very much sir.
@@BabinRoy-i9e Thank you for your kind words. I am glad you find the lectures useful and enjoyable.