No - Ferroelectric materials do not lose their polarization just by going through the phase transition temperature. The P vs. T diagram at high temperature above the transition makes the material Paraelectric. At the transition temperature down to a much lower temperature, it becomes ferroelectric. This is analogous to magnetization versus temperature. Due to the Devonshire-Landau-Ginzburg theory, the phase transition temperature is off a bit from the Currie temperature. Now, polarization comes also from applying a field - thus, this can be where trouble can start: In thin films, the surface may be so defective that the domains can no longer switch within a few polarization reversals. Then, you have no way to show that the material is ferroelectric unless you can a pply a much higher field and de-pin the domains. This is not a cure all process, because domains do not de-pin completely and then the "fatigue" arrives even fast. The sheet of charge due to domain pinning is strong enough to screen the outside field. And, in this case the material seems to have lost its ferroelectricity. 

When a ferroelectric material is poled, we are essentially applying potential to reorient the domains to align in a particular direction. When heated beyond the Curie (or depolarization) temperature, the structure essentially becomes centrosymmetric and all polarization is lost. However, once it cools down to a level below Curie temperature the non-centrosymmetry is re-established, albiet in a fashion where domains are now aligned randomly. So in a sense, yes, bulk polarization will be lost once a poled material transcends the Curie Temperature and it might be required to to pole the sample again, for all practical purposes. 

For thin films, the transition temperature is smeared by surface space charge. So, the P vs. T diagram is always completely second  order. Thus, it has a long tail. The cause of this can be related to a depolarizing field that reaches a maximum near the surface and then goes to zero. Thus, we have an inversion region, which makes the depolarization P vs E curve completely different from bulk. The E versus E  curve is more like a complementary error function, peaking at the Bulk value of the Polarization. This effect of thin-film inversion is simply because it is more important the internal depolarization as the electrode is screening the polarization. Thus, a combination of non-stoichiometry (defects) and the overall local surface field, which is proportional to the gradient of the polarization  plus defect gradient, and  this field in the reverse direction, makes the complete destruction of the Bulk polarization after transversing through a now re-defined Transition temperature not possible. All FeRAMs,require this inversion property to work as a storage device. Therefore, the textbook perfect answer By Aditya Chauhan is completely true for bulk ceramics, where the surface inversion is not in the same proportion as the overall bulk capacitance. Unfortunately such concepts fail in thin films. The evidence of this field is well shown in the literature for thin films and FeRAM.

In general I am agree with prof De Araujo. After the heating above nthe curie point you lose the ferroelectric behavior, but in th cooling process local ferroelectric properties appears again.. You can observe that usin Raman spectr. or permittiy versus Temperature

After heating above the Curie temperature ferroelectric properties disappear

but on cooling the ferroelectricity appears again if stoichiometry of the sample preseved. Decomposition at high temperatures may be a reason of the sample degradation and disappearence of the ferroelectric properties.  Verify your sample after heating-cooling runs.

The ferroelectric properties will reappear but not the for the poled ones. It is like a permanant magnet. If you heat it above Tc it will be still magnetic but with out any poles. You got to remagnetise them. It is the same with ferroelectric too.