Q 1) How will you
choose the right algorithm?
Using past data and optimize future
result is called as machine learning. There is various techniques using which
we can select the right algorithm.
Depending upon the two things the
algorithm is choose. First is Data and second one is target value.
If you have target value then we will
select the supervised learning.
If your target value is like A/B/C or
Yes/No then we will select “classification”.
And if the target value is numeric then
we will go far the “regression”.
And if you don’t have any target value
then go far the unsupervised learning i.e “clustering”. Clustering means
putting the similar type of data together.
Q 2)Write in detail
about the different steps that you will perform to develop the m/c learning
algorithm for a given scenarios?
There are following
steps which is required to develop a m/c learning application.
1)Collect a data.
2)Prepare a i/p data.
3)Analyze a i/p data.
4)Train the algorithm.
5)Test the algorithm.
For spam detection we
choose decision tree algorithm. Decision tree algo. Build a tree structure base
on the given training set of data, which is used to classify unlabeled data.
for spam detection we use itrative
dichotomiser 3 (ID3) algorithm.
This algo. Build a decision tree based on entropy and the information gain.
Automated breaking system is used in self-driving cars. There are three major task that need to be
consider while developing the algorithm the are as follows:
The detection of an
The Identification of
Localization and Prediction of Movement of the object.
The m/c learning algo. Classify this above
steps in to following classes. decision
matrix algorithms, cluster algorithms, pattern recognition algorithms and
Stock market prediction
For the stock-market prediction the regression technique is
used. Suppose we want to compare the bank nifty price affect the canara’s stock
price. In this case the algo. Will take the first 40 days data as a tanning set
and last 20 days data as a test data.