Vector |
Structured Learning |
Unstructured Learning |
Reinforced Learning |
What |
Needs Humans
Learn relationships
Needs explicit output
|
explore data
no explicit output
|
perform a task
get rewards
|
When to use |
Classify data - humans
define type of behaviour
algorithms form new data sets
|
No data classification
Finds patterns with data set
reclassification data to form new data
|
No Training data
NO end state
Need to interact with model to derive output
|
How to use |
Label all input data
define output
find connections between data sets
apply algos to form new data sets
|
Unlabeled data
infers structures from data
ID's groups of data with same behaviour
|
Takes an action/task
gets a reward optimzation based on best series of actions
|
Various Algo's and business case
1 |
Liner regression
Highly interuptable . need to have past relationship identified
for the output variable
helps predicate future output variables
|
understand sales drivers such as price, ad's or distribution
optimzation of price points
|
2 |
Logistic regression
used for classifation tasks, meaning the output variable is binary (eg, only black or white)
rather than continuous (eg, an infinite list of potential colors)
|
Predict if a skin lesion is benign or malignant based on its characteristics (size, shape, color, etc) |
3 |
Linear/quadratic discriminant analysis
a logistic regression to deal with nonlinear problems—those in which changes to the value of input variables do not result in proportional changes to the output variables
|
Predict client churn
Predict a sales lead’s likelihood of closing
|
4 |
Decision tree |
|
5 |
Naive Bayes |
|
6 |
Support vector machine |
|
7 |
Random forest |
|
8 |
AdaBoost |
|
9 |
Gradient-boosting trees |
|
10 |
Simple neural network |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Comments (0)
You don't have permission to comment on this page.