Predictive Analytics or Predictive Modelling, the most prevalent form of data mining, use advanced statistical techniques or machine learning models to predict the probable outcomes. Statistical techniques apply various mathematical concepts to data and derive inference and insights. Hypothesis and assumptions form the core of statistics in which relevant knowledge particularly human expertise is required in formulating and testing the hypothesis thereby predicting the outcomes.
On the other hand, machine learning uses reverse approach to predict the outcomes. Machine learning predicts the outcomes first and based on these outcomes it determines the model that uncovers the hidden driving factors of the outcome. Various complex and non-linear relationships can be uncovered from the machine learning models. Large number of variable relationships and interactions can be identified. The models adjust and improve over the time. The factors such as the above emphasize the importance of machine learning in making predictions.
Machine Learning learns automatically with algorithms. Based on learning, it can be broadly classified into three categories:
1. Supervised Learning
2. Unsupervised Learning
3. Reinforced Learning
The other classification of machine learning can be based on the types of desired outcomes or the type of problem solving. The common types of problem solving are:
1. Classification
2. Regression
3. Clustering
4. Rule-Extraction
Sign up here with your email
EmoticonEmoticon