A nova occurs when matter accretes on the surface of a white dwarf from a nearby binary companion and ignites in a tremendous thermonuclear explosion. Often the phenomena is confused with a supernova, although if enough matter accumulates on a white dwarf for the star to exceed the Chandrasekhar limit, a type Ia supernova will occur instead. Also unlike a supernova, which completely destroys the star, a nova is capable of repeating under the right conditions. 
A nova is categorized in three main varieties: classical novae, recurrent novae, and dwarf novae. The first two categories depends on how often the nova occurs, while the last is different through its behavior and source of the nova outburst.
Classical novae are classified by the sudden increase in brightness by a order of magnitude between 103 and 106 of a white dwarf star and the corresponding sudden ejection of surface matter from the that same star. This is caused when a white dwarf star shares a close orbit with a binary companion star that is typically in its red giant stage of its stellar evolution. As the companion expands as a red giant, its outer surface of hydrogen exceeds its Roche-lobe and is gravitationally drawn from its surface towards the white dwarf, forming an accretion disk over time before being pulled down to the surface of the white dwarf itself. As the mostly hydrogen gas builds up on the degenerate surface of the white dwarf, it begins to compress from the accumulation, causing the gas and the surface underneath to heat up. Once the temperature of 10 million K is reached at the point the overlying hydrogen shell and original surface meet, the hydrogen shell suddenly ignites and begins to fuse the hydrogen into helium. This in turn causes a runaway thermonuclear process that results in the sudden increase in brightness (magnitude increase of 8 to 15), and the loss of the surface material as it explodes away as a fast expanding shell.
Classical novae are in three sub-categories based on their light curves:
Fast novae reach their maximum brightness from the explosion very quickly, but only maintain this magnitude for a few days before a steep decline in their luminosity occurs. Usually in three months time, the star has faded to only a tenth of the original maximum luminosity caused by the initial outburst.
Slow novae have a more gradual rise to maximum brightness and will maintain that magnitude of luminosity for several weeks to mouths before declining. The gradual fading will first be relatively slow and will fluctuate, with the rate of decline increasing over time. The star often will brighten again to a second maximum of luminosity before fading back to the original low luminosity. It usually takes around 5 months for the star to fade to one-tenth of the original maximum luminosity.
Very Slow Novae
Such novae are often referred to as symbiotic novae or RR Telescopii stars. These novae will often maintain the maximum luminosity from the explosion over several years. As such it may take decades for the luminosity to decline back towards the star's original brightness. One star that was observed as a very slow novae was RT Serpentis in 1915. It slowly increased in luminosity and maintained its maximum for 10 years before declining.
Recurrent novae are much like classical novae but are different as they are observed to have more then one outburst. The terms classical novae and recurrent novae may be interchangeable, as classical novae may have more than one outburst, given enough time.
The amount of time between eruptions of a novae is determined by the original mass of the white dwarf itself, and thus the amount of accreted material required to trigger the outburst. Binary stars with a white dwarf that has a mass of only 0.6 solar masses may take as long as 5 million years to accumulate enough material from its red giant companion before exploding as a nova. A white dwarf that is 1.3 solar masses on the other hand, may only take 30,000 years before accreting enough matter to have an outburst, having such eruptions every 30,000 years.
If a heavy white dwarf manages to accumulate enough additional mass to surpass its Chandrasekhar limit (1.4 solar masses) before the gas on the surface heats up sufficiently to cause a nova, the white dwarf itself will begin to collapse from the increased gravity and perish in a type-Ia supernova explosion instead.
Dwarf novae, also known as U Geminorum-type variable stars, are a type of cataclysmic binary that have occasional, abrupt increases of brightness between 2 and 6 magnitudes.
Normally, the observed light comes from four distinct sources in the binary system: the white dwarf primary, the cooler secondary, the accretion disk, and a hot spot. This result occurs when a binary pair (a white dwarf primary and its secondary star, typically a subgiant) have a mass between 0.5 and 1 solar masses and are locked in a very close orbit (the orbit taking only between 3 and 15 hours). As they orbit, matter from the companion collects around the white dwarf in an accretion disk forming a hot spot. From the point of an observer this gives the appearance of a single source of light.
Although U Geminorum stars typically have routine small variations in light over time. it is the unpredictable, sudden increase in order of magnitude that also classify these stars as dwarf novae. There are two competing theories at present to explain these outbursts. The first, called the mass-transfer burst model is where a sudden increase of material is transferred from the secondary companion to the white dwarf, causing the accretion disk to collapse, dumping a large amount of hydrogen on the surface of the white dwarf. It is thought that this tremendous release of energy through the sudden mass-transfer is the source of the outburst and corresponding sudden rise in magnitude. The other theory, called the disk-instability model, maintains that gas from the secondary subgiant builds up on the outer edges of the accretion disk around the white dwarf. When critical density occurs on this outer edge, thermal instabilities within the disk causes the matter to collapse onto the surface of the white dwarf itself. This release of energy from the accretion disk's instability releases gravitational energy, which is the source of the sudden corresponding rise in magnitude.