With many stars being born every day, and many more dieing and disappearing, the space is a massive challenge to deal with. Today, thanks to digital photography astronomers get to spend more time making sense of the sky, than just gazing at it with telescope every night.  The best thing that astronomer say, is the use of data mining and statistic techniques, which makes their job a lot easier.

Some key facts:

• Key goal is learning about the changing sky faster and more efficiently.
• Typically, 1.5 Million new observations every night.
• Machine learning algorithms which recognize different galaxy types ranging from spiral to elliptical are improving and aiding the process very efficiently.
• Its also important to note that more data also means increased problem space.

[ Read  more at Space]

The data sent by spacecrafts during Space Missions, is it small or big?

There are tremendous amount of information and data sent by space missions, imagine the sounds send by voyager, but in truth there is more data being sent than just audio, video and images. The data of space missions also consist of the actual payload data and housekeeping telemetry about the status of its various systems. These data are under constant analysis for the investigation of causes of occurred anomalies. Housekeeping parameters are used to check the health of systems and sub-systems and to ensure that those systems are not facing any problems like rise in temperature or an increase in power consumption of a particular system.

Where does data mining come into play?

A data mining technique called Distance-Based Anomaly Detection (DBAD) comes into action whenever some problems when occurred affects the parameters of a system which are not normally associated with that situation. The idea behind DBAD is to use vectors which can be treated as points in an N-dimensional vector space. These vectors can be formed by combining N different housekeeping parameters. The Euclidean distance metric (or some other suitable metric) can then be employed to calculate the distance between different points.

What procedures can be applied for measuring reliability using DM?

Initially, normalization of the different parameter (like temperature, pressure and power) values is done because dissolve difference in units. {One method of normalization is to use the mean and standard deviation calculated for each parameter then subtracting the mean from the current value and dividing the difference by the standard deviation}. In addition, some weight is applied to each parameter according to its significance. After completion of the above step, clustering of the vectored points in certain regions of the N-dimensional space may start. The presence of an anomaly is doubted by some out of range values in the observed data sets indicated by the points lying outside the clustered regions. The anomaly is not confirmed because such values can be the result of some planned events like the firing of an engine to change the orbit of the satellite etc. There are more techniques available to investigate space related data and ensure a reliable system in future. It is very important for preventing some predicted happenings.

[Read related works]