Clustering Machine Learning
This unsupervised machine learning trains models without pre-knowledge of group identity for large amount of observations. The training algorithms automatically compute data to group observations into predefined number of groups to achieve least distances within groups and maximum distances from different groups. Observations within a group have the similar characteristics. Clustering machine learning algorithms are available in many statistical analysis software such SAS, Python, R, and Matlab. The methods are widely used in targeted marketing and medical research.
We streamline the cluttering machine learning in our analysis process from data loading and preparation to model building and evaluation, and results output for further analysis and business strategy development.
Predictive Machine Learning
This supervised machine learning trains models based on known outcome (dependent variable) and features (independent variables) possibly effecting to the outcome. There are linear and non-linear supervised machine learning algorithms which handle different types of dependent variables. Regression modeling is a type of linear modeling that is widely used in many industries to find what features (independent variables) impact on outcome. Examples of supervised machine learning are targeted marketing, medical treatment effectiveness analysis, product demand forecasting, equipment breakdown prediction, fraud detection, and so on. Supervised machine learning algorithms are available in many statistical analysis software like SAS, Python, R, and Matlab.
We provide services in predictive machine learning to analyze data in depth and causality way, which greatly elevate data power in business strategies and operations. We streamline the process from data preparation to modeling, and business application, which make the predictive analysis feasible, affordable, and accountable.
The example applications of predictive machine learning are sales prediction, targeted promotion, fraud detection, defect reduction, and so on.
Please contact us for potential projects.
Bayesian Probability Analysis
In real world, many things happening are due to certain conditions. Bayesian theorem is used in computational algorithms to calculate the probabilities of one event happening under certain given condition. In recent decades, Bayesian probability has been applied in many business areas such as targeted marketing and medical research.
Bayesian probability machine learning is available in Python package. We have developed algorithms to compute conditional probabilities for business applications.
