Cyberiafreak

"Fortunate are those who take the first steps.” ― Paulo Coelho

ML : How to develop a mode, algo use and one liner explanation

Supervised Learning – Classification, Regression

Unsupervised Learning – Clustering, Recommendation

————————————————————————————————-

(1)Classification (Y/N/not sure, Maligant tumour or not)
==================
Which customer are more likely to buy, stay, leave(churn analysis)
Which customer is going to get loans
Which transactions are fradulent
Which quotes are more likely to get executed

   What are the algorithms are there for Classification..
1.BCD (Boosted decision tree)
2.DF (Decision Forest)
3.DJ (Decision Jungle)
4.LR (Logistic Regression)
5.SVM,ANN

(2) Regression (Predictive analysis)
==============================
1. Stock price prediction
2. Market forecast
3. Weather forecast
4. Calculate insurance premiums based on different factors(feature)
5. Manpower prediction
6. Workload prediction
What are the algorithms are there for Regression..
1. Bayesian Linear
2. Linear Regression
3. Ordinal Regression
4. (BCD) ANN, Boosted Decision Tree
5. DF (Decision Forest)

(3) Cluster (Segmentation)
==============================
Divide a customer base into groups of individuals, age, market segmentation,
Data compression, color reduction, pattern recognition
             Which Algorithm to use..
1. K-means

(4) Recommendations (Product recommendation)

=======================================================
1. Amazon Recommender system
Algorithm: Matchbox

Advertisements

August 11, 2014 Posted by | Azure ML, MarketLearning | Leave a comment

   

%d bloggers like this: