The security in the interpretation of spectra can only learned with time, it requires experience.
You should know the essential spectral classes and their characteristic spectral properties (read the Kaler, ISBN-13: 978-0521585705).
You should also know as many lines. Practice that very well in VSpec! In this program you will find a tool that lets you show the periodic table and the lines of the elements and ions can fade in the measured spectrum. So you can quickly and easily check whether your assumptions could agree. But beware: The software displays options. If for example a HeII line can occur in the spectrum of a star who does not know VSpec. You should know this. A HeII line is not appearing in the optical spectrum of a G-Star. It’s just not hot enough to produce HeII too. He has an extremely high ionization energy. HeII lines are therefore in the spectra of hot objects, such as O-stars. Therefore read the Kaler!
Often we find diffuse lines (DIB’s) of interstellar matter (ISM). They do not belong to the object star, but are ‚on the way‘ absorbed from the starlight. The nature of the ISM is still partially unknown. It is probably the finest dust particles, which bear on their surface absorbing molecules, such as polycyclic aromatic hydrocarbons (PAHs).
You should immediately recognize the telluric lines, especially the O2 bands and water lines (in case of high resolution spectroscopy). Sometimes you can orient yourself on such bands in the spectrum, if the grating position or wavelength window is not accurately known. Over time you develop an eye for it.
For practice please interpret the adjacent spectrum of Mizar A (dispersion 0.45 angstrom / pixel, resolution about 1.5 angstroms). The solution to this find here. How it works? Are you fit in interpreting?
You spectroscopy a star with a slit spektrograph in the wavelength range around the H alpha line and get the image reproduced below – and wonder. What is that? The rather uniform spectral continuum of that star is obscured by the bright transverse stripes?
This is not an artifact! You have spectrocopied a star in the line of sight of an emission nebula (HII region). That are emission lines of the nebula, the intense vertical line is the picture of the slit in the light of the H alpha line in emission.
It is theta1 Orionis C, the brightest star in the Trapezium, a very hot and rapidly rotating young O star with a magnetic field (magnetic rotator). The highly luminous star in the UV is responsible for the HII region in the Orion Nebula (Ionization). It lies close to the Orion Molecular Cloud and erodes it with its bright light. If you look closely, you will find in the H alpha emission line in the top quarter of the nebula, a diffuse broadening of the line to short-wave contour. Obviously, moving here a part of the emitting hydrogen atoms toward us.
An example of successful collaboration between amateurs and professional astronomers observing that interesting star system you find in 0805.0701v2. So amateur spectroscopy is fun (and hard work)!
Below the evaluated range of theta1Ori C (not rectified). R = 12000.
The calibration was carried out using 4 water lines, each pair of left and right of the H alpha line, so that the wavelength scales accuracy should be in the area of H alpha line better than 0.05 Angstroms. The wavelength scale thus refers to the reference frame of the observer, is therefore not heliocentric or barycentric corrected.
The rapidly rotating O star has a rotational broadened H alpha line. Because the nebular emission, the minimum of the absorption line is hardly determinable. It may well also be below the emission line.
In the calibrated range can now identified the other emission lines: 6548 and 6584 are prohibited NII lines, 6678 HeI. The peak at 6649 is an artifact. The calibration was done with two water lines on both sides of the H alpha line (because I’m only interested in the H alpha). The calibration function was extrapolated, which is why it is incorrect to the right edge of the spectrum.Therefore, we find the HeI line appears at 6685 angstroms, instead of the correct 6678.
To identify unknown lines this table is helpful: download here.
Astrospectroscopy 