Holiday Ayo - The use of artificial intelligence (AI) is not only limited to the technology sector, but apparently can be relied on in space research.

The National Research and Innovation Agency (BRIN) revealed that AI technology can be used for research in the space sector, to look for something new. 

“As space researchers, of course we hope to be able to utilize (AI technology) to complete research so that it is more up-to-date," said Head of the BRIN Space Research Center, Emanuel Sungging, quoted in a written statement, Tuesday (16/7). 

Meanwhile, Lecturer at the Department of Computer Science and Electronics, Faculty of Mathematics and Natural Sciences (FMIPA), Gadjah Mada University (UGM) Edi Winarko in his presentation focused on the application of no-code or without using code in machine learning (ML). 

“Today's ML users are not only people with a computer education background, so we have developed tools that are easy to use for modeling, thus democratizing AI, so that it can be used to help and make it easier to use ML for everyday work, including space research," explained Edi. 

In the ML-based research cycle, there are several stages, including identifying the problem or problem formulation, data collection and data selection, selecting an appropriate algorithm, evaluating model performance, and operationalizing the model. 

“There are several methods for collecting data that can be done, including downloading public data, using APIs (Twitter API, WikipediaAPI, WeatherAPI for weather data), and finally downloading via web scraping either with full coding or no-code," said Eddie.

To have quality data, it is necessary to carry out data processing such as data cleaning to handle missing values and outliers, data transformation through normalization, standardization, discretization, and also data reduction after features and feature creation.

Edi then explained two ML tools that can be used for space research, including Orange Data Mining and PyCaret. 

Orange Data Mining is an open source based data science tool used for modeling analysis of various types of data, tabular, document, time-series, graph (network), and images, by creating workflows (connecting widgets to one another). 

While PyCaret is an open source low-code data science tool that covers all stages in the ML project lifecycle and can be combined with Tableau, Power BI.