LIMITAIONS OF THE SYSTEM
1.Limitations of the recognition process:
The system has the following recognition limitations -
• We are storing the commonly found verbs only 500. Though it is very unlikely that ordinary Sanskrit literature will overshoot this list, yet the system is likely to start processing a verb as subanta if not found in the database.
• We have short verb list without added upsargas with verb roots and secondary suffixes if any verb come with upasargas or secondary verbs then system fail to recognition as verb.
• This work assumes initial sandhi processing, without which, some results may turn out to be incorrect because we have verb and avyaya list only main form in our database. In this condition system fail to recognize correctly as verbs or avyayas and recognize as subanta then send it for analysis. For the example: upaagacCati (upa + aagacCati)
• We have 519 avyaya lists in our avyaya database; it’s not enough for avayaya recognition in ordinary Sanskrit literature. In this condition the system is likely to start processing an avyayas as subanta, if it’s not found in avyaya database.
• Some forms ending of primary affixes looking like subanta but it’s originally avyayas. For the example: tumun, gattvaa, aadaaya, aagatya etc. System will recognize as subanta of these type avyayas and send for analysis and system will analyze wrong.
• Many subantas (basically shtRu pratyayaant in locative singular number) are looking like verbs, for these subantas system will be recognize as verb, so system can’t analysis these types subantas. For the example: bhavati, gacCati, pathati, chalati. etc.
2. LIMITAIONS OF THE ANALYSIS PROCESS:
The system has the following analysis limitations -
• Same forms are available in the dual of nominative and accusative cases like- rmau, dual of instrumental, dative and ablative cases like- rmbhym, plural of dative and ablative cases like- rmebhy, dual of relative and locative cases like - rmayayo. In neutar gender as well, the nominative and accusative singular forms may be identical as in pustakam (1-1 and 2-1). In such cases, [10] the system will give all possible results as in
rmau = au (1.2 & 2.2)
rmbhym = bhym (3.2, 4.2 & 5.2)
rmebhy = bhyas (4.2 & 5.2)
rmayayo = os (6.2 & 7.2)
pustakam = su (1.1 & 2.1)
hare = as (5.1 & 6.1)
• Some k•danta forms (generally lyap, tumun, and ktv suffix ending) look like subanta (for example - vihasya vihya, dya, gtv, pathitv etc.). In such cases, the system may give wrong results like
vihasya = viha + as [(6.1) (masculine ‘a’ ending )]
pathitum = pathitu + am [(2.1) (masculine ‘u’ ending )]
gantum = gantu + am [(2.1) (masculine ‘u’ ending )]
netum = netu + am [(2.1) (masculine ‘u’ ending )]
pradya = prada + e [(4.1) (masculine ‘a’ ending )]
vihya = viha + e [(4.1) (masculine ‘a’ ending )]
dya = da + e [(4.1) (masculine ‘a’ ending )]
gtv= gtv + su [(1.1) ( feminine ‘’ending )]
chalitv = chalitv + su [(1.1) (feminine ‘’ ending )]
pathitv = pathatv + su [(1.1) (feminine ‘’ ending )]
to solve this problem, we are trying to store these k•danta forms of the 500 commonly found verb roots.
• Same forms are available in the dual of nominative and accusative cases like- rmau, dual of instrumental, dative and ablative cases like- rmbhym, plural of dative and ablative cases like- rmebhy, dual of relative and locative cases like - rmayayo. In neutar gender as well, the nominative and accusative singular forms may be identical as in pustakam (1-1 and 2-1). In such cases, the system will give all possible results as in
rmau = au (1.2 & 2.2)
rmbhym = bhym (3.2, 4.2 & 5.2)
rmebhy = bhyas (4.2 & 5.2)
rmayayo = os (6.2 & 7.2)
pustakam = su (1.1 & 2.1)
hare = as (5.1 & 6.1)
• Our system has not any gender marker right now, so system can’t give accurate gender information. Currently our centre is working on the gender information also, when it works will complete then we will add with this system and our system will give correct gender information also.
• Same characters are found in the ending of some subanta-padas, so system can’t recognize correctly. So this system is also weak in accurate praatipadikas information in some cases, who is same characters ending. Like ‘h’ ending and ‘sh’ ending praatipadika’s last character change in ‘T’ and ‘D’. So system can’t recognize correct praatipadikas. In this condition system will give more possible result. For the example: liD, liT, viT, viD, taadRuk, taadrug etc. in this cases system will give result:
liD = lih + su / lish + su [prathamaa ekavacana]
liT = lih + su / lish + su [prathamaa ekavacana]
viT = vih + su / vish + su [prathamaa ekavacana]
viD = vih + su / vish + su [prathamaa ekavacana]
taadruk = taadruh + su / taadrush + su [prathamaa ekavacana]
taadrug = taadruh + su / taadrush + su [prathamaa ekavacana]
• Same forms are available in the some vibhaktis. So this system can’t recognize crroect suffixes. In this condition system will print more possible results.
rmau = au (1.2 & 2.2)
rmbhym = bhym (3.2, 4.2 & 5.2)
rmebhy = bhyas (4.2 & 5.2)
rmayayo = os (6.2 & 7.2)
pustakam = su (1.1 & 2.1)
hare = as (5.1 & 6.1)