Machine learning

AI-based machine learning is commonly used to create prediction models. Qlucore Omics Explorer is the powerful visualization-based data analysis tool with inbuilt powerful statistics that delivers immediate results and provides instant exploration and visualization. The program supports a broad spectrum of Omics and NGS data. The program also supports the building and use of prediction models.

User friendly machine learning

With the inbuilt functionality you can approach machine learning in a fast and efficent way. Machine learning started as a research field in the 1950s. The techniques have since been applied to a wide range of applications including computer vision, video games, adaptive websites and speech and handwriting recognition.

Real world data is often challenging in several ways. The problems are present both when building a classifier, and with applying it. Common errors includes error in wet lab work, incorrect annotations (applicable for building a classifier), incorrect output from instruments (i.e. pure errors), missing data, noise etc.

Qlucore Omics Explorer includes several different machine learning techniques such as:

Boosted trees
Support vector machines (SVM)
Random trees (RT) and
Nearest neighbours (kNN).

Building a classifier (predictor)

classification The program Qlucore Omics Explorer includes several different machine learning techniqueshat can be used to build a cl

assifer. Classifiers can be built utilizing an internal cross validation scheme or by using an external validation data set.

The quality of data and annotations will be expected to play a significant role on how good classifiers can be built, and that is why the Qlucore visual approach is so important. Inspecting data and securing quality for building models is key to securing quality output.

Visualization and dimension reduction techniques such as principal component analysis (PCA) are especially useful since it gives an overview and assist the user to focus on key structures. Outliers are very often also easily detected. Knowledge about data and the specific application is very valuable since data inspections is best performed by the user groups who have the domain knowledge. These user groups do not only work with data analysis and hence is the requirement for user friendliness is very important.

Read more about How to work with classifiers.

Solutions in addition to Qlucore Omics Explorer

The machine learning functionality is integrated in all our solutions. For clinical data we provide dedicated models for different disease areas such as leukemia, lung cancer and bladder cancer. Read more here.