Top Page | Upper Page | Contents | About This Site | JAPANESE

Materials Informatics and Chemometrics

Materials informatics and chemometrics both use Data Science for chemical investigation and research.

Experiments are important in the field of chemistry, but they take time and money. The money includes the cost of properly disposing of the substances created in the experiment so that they do not affect the human body or the environment. Some experiments are dangerous, so we want to reduce the number and amount of experiments as much as possible.

Materials informatics and chemometrics have such needs.

Why multivariate analysis is useful for chemical research

For example, when identifying the type of chemical substance, it is measured using various measurement methods. A single measurement method may output multiple values. Identification is possible by using data from multiple perspectives.

Also, when trying to express chemical substances, there are numbers and types of atoms, shapes of molecules, etc., so multiple parameters are used.

In either case, one chemical substance is represented by many items. From a data science perspective, this is a "problem with many variables (multivariate/multidimensional)". Therefore, multivariate analysis is useful.

Materials Informatics

In materials informatics, we use a lot of experimental data to develop unknown substances (materials). For materials and experimental conditions that have never been tested, we will predict experimental results from past experimental data.

Multi-Variable Analysis and Artificial Intelligence (AI) are used for this prediction .


Chemometrics applies data science to the field of analytical chemistry. Like materials informatics, we also search for unknown substances.

We als

Structure activity relationship

Structure-activity relationships are a method for coming up with new drugs. Investigate and utilize the correlation between the molecular structure of a compound and the strength of its bioactivity . Since the molecular structure is represented by multiple variables, multiple regression analysis is used.

Although it is called "structure-activity relationship," it can also be used to analyze physical properties by changing the target variable for activity to physical properties. Structure-activity relationship is a method that does not appear in chemometrics, but if you change the activity part to physical properties, it becomes almost the same as chemometrics.

Difference between materials informatics and chemometrics

Many books on chemometrics were published in the 1990s and early 2000s.

Materials informatics (MI) has an aspect in which a part of chemometrics, which utilizes experimental data, has made great progress with the power of the latest computers.

Materials informatics has become a field in the position that "there is already information (data)". In addition to the measurement data of substances, the use of chemical-related data is also spreading.

On the other hand, chemometrics considers not only the existing data, but also the management of measuring instruments, the scene of obtaining data, such as Design of Experiments , and how the data was measured. It also includes processing data.

NEXT Chaos