Assignment1 Instructions 2017

Assignment1_instructions_2017

Assignment1_instructions_2017

Assignment1_instructions_2017

User Manual: Pdf

Open the PDF directly: View PDF .
Page Count: 1

Human-centered Assistive MACHINE LEARNING

Robotics IN ROBOTICS

Technische Universit¨at M¨unchen Assignment1 Instructions

Prof. Dr.-Ing. Dongheui Lee

SUBMISSION

Each student must work independently. Please upload a file called Assignment1Surname ID.zip (where

Surname is the surname of the student that submits the file and ID is the enrollment number) on moodle.

This file should contain:

•Assignment1Surname.pdf, a pdf file containing the solution to all the exercises (see below for

further informations)

•The Matlab code in a subfolder called Code. Students can create any functions that they consider

necessary to solve the problems.

The submission deadline is on the 23.06.2017 at 23.59.

HW1 Surname.pdf.

•Students need to provide a pdf file containing the solution to all the exercises. Students must clearly

indicate in this file to which exercises and to which question the solutions refer to.

•For Exercise1.a −b)report learned parameter values as well as optimal values of p1and p2for

k= 2 and k= 5.

•For Exercise1.c)attach the required plots.

•For Exercise2.)report the optimal dvalue, its classification error and confusion matrix. Also attach

a plot of classification errors when varying dfrom one to sixty.

•For Exercise3.a −b)attach the resulting plots.

Subfolder Code.

•For Exercise1provide a matlab function Exercise1.m. The input to this function is kand it’s

output is the cell array par.

•For Exercise2provide a matlab function Exercise2.m. The input to this function should be dmax

which is 60 in this exercise. The outputs of this function should be a plot of classification errors (from

d= 1 to dmax ), optimal value of dand its classification error and the confusion matrix.

•For Exercise3provide a matlab functions Exercise3kmeans.m and Exercise3nubs.m. The inputs

to Exercise3kmeans.m are the motion data, the initial cluster label and the number of clusters.

You can’t use the matlab function “kmeans”. The inputs to Exercise3nubs.m are the motion data

and the number of clusters. The outputs of both function are the 3plots required in Exercise3.a−b).

NOTES

•Do not include the provided datasets in you submission.

•We will mark the assignment even if you forget to press the submit button.