This dissertation is an effort in adding to existing methods of predictive capability for space weather events such as coronal mass ejections (CMEs) and solar flares and to predict if these events will cause a geo-magnetic storm well in advance. I use solar observational features – filaments and sigmoids, and atmospheric helioseismology - as tools to identify if CMEs will cause disturbances in the magnetosphere, and when a solar flare might occur, respectively. The main elements of this dissertation are as follows. First, I study the relation between axial field directions of filaments on the Sun to that of their interplanetary CME (ICME) counterparts near Earth. Filaments are plasma fibrils aligned along local magnetic fields and always form over polarity inversion lines. These features often erupt as CMEs and filament material have been observed in CMEs. I show, using an analysis of about two solar cycle filaments and ICME data, that the magnetic orientation of the filament in a CME is retained during its traversal of interplanetary space. I find that for solar cycle 23 and 24, the field orientation is retained 85\% and 65\% of the times, respectively. Second, I compare chiralities of filaments associated with their X-ray sigmoid counterparts to identify if there is consistency between the two. Non-linear force-free field surface calculations are performed using vector-magnetograms to compare the calculated helicity sign of sigmoids with that of the observed signs of filaments and the handedness of sigmoids. I find that the sigmoids for which the signs from the three modalities do not match, produce 3.5 times as many flares as that when they do match. Third, I use Magneto-Optical filters at Two Heights (MOTH) Dopplergrams in KI and Na-D2 lines with continuous observations of the Sun covering non-flaring and flaring times. Using Fourier cross-correlation and phase-difference techniques, I observe changes in travel-time for waves close to the acoustic cut-off frequency. At 6.5mHz frequency, large travel-time deviations from the mean, almost 3-15 hours prior to the occurrence of flares are observed. The results suggest that the methods utilized in this work have a potential in improving space weather predictions.