Jese jese dunya bare data ke daur main dakhil hue, to isko zakheera karne ki zaroorat bhi barh gae. Aaj hum is artile main parhenge ke Data Science kya hai?
2010 tak enterprise industries ke liye ye ek bunyadi challenge aur tafteesh thee. Us waqt bunyadi tawajah data ko zakheera karne ke liye ek framework aur iske hal ko par thee.
Ab jab ke Hadoop aur doosre frameworks ne storage ke masaile ko kamyabi ke sath hal kiya hai to, tawajah us data ke processing ki taraf muntaqil ho gae hai.
Saare khayalat jo aap Hollywood ki sci-fi movies main dekhte hain, wo Data Science ke zariye se haqeeqat main badle jaate hain.
Data Science Artificial Intelligence matlab ke Masnui Zahanat ka mustaqbil hai.
Lihaza, ye janna buhut zaroori hai ke Data Science Kya Hai? aur ye aap ke karoobar ki qadar ko kis tarah se barha sakti hai?
Data Science Kya Hai?
Data Science mukhtalif tools, algorithms aur machine learning ke usoolon ka ek murakab hai, jis ka maqsad khaam data se posheeda namoone daryaft karna hai. Lekin ye is se mukhtalif kese hai jo statistics ke maahreen barson se karte aa rahe hain?
Iska jawab (Difference between prediction and explaination) matlab ke wazahat aur peshan goi ke darmiyan faraq main hai.
Doosri taraf Data Scientist na sirf us se baseerat matlab samajh boojh daryaft karne ke liye tajziya karta hai, balke mustaqbil main kisi khas waqiye ki mojoodgi ki shanaakht ke liye mukhtalif jadeed machine learning algorithm bhi istaimaal karta hai.
Data Scientist data ko kae zaawiyon se dekhta hai, kabhi kabhi ese zaawiye jo pehle kisi ko maaloom nahi the.
Lihaza, Data Science bunyadi tor par faisle aur pesh goeyan karne ke liye istaimaal kiya jata hai, jo ke pesh goi karne wale asbab ke mutailq tajziya, nuskhe wale tajzye (pesh goi ke ilawa faisla karne ki science) aur machine learning ka istaimaal karte hain.
- Pesh goi ke sabab wale tajziyat (Predictive causal analytics)- Agar aap ek esa model chahte hain jo mustaqbil kisi khas event ke imkanat ki pesh goi kar sake, to aap ko faisla karne ki science par amal karni paregi. Agar aap udhar raqam dete hain, to mustaqbil main credit ki adaigi waqt par karne wale saarfeen ka imkan aap ke liye tashweesh ka baais hai. Yahan aap ek esa model bana sakte hain jo gaahak ki adaaigi ki taareekh par pesh goi karne waale tajziyat anjaam de sake taake ye andaza lagaya ja sake ke mustaqbil ki adaaigi waqt par hogi ya nahi.
- Wazaahati Tajzyat (Prescriptive analytics) - Agar aap ek esa model chahte hain jis main apne faisle karne ki zahaanat mojood ho aur usay mutahrik parameters ke sath tabdeel karne ki salahiyat ho, to aap ko yaqeeni tor par us ke liye nuskhaati tajzyat ki zaroorat hogi. Ye nibatan naya shoba mashwaray faraaham karne ke baare main hai. Doosri sharaait main, ye na sirf pesh goi karta hai balke tajweez karda aimaal aur mutailiqa nataaij ki ek had tajweez karta hai. Iski behtreen misaal google ki self-driving car hai. Gariyon ke zariye ikhata kiya gaya data self-driving cars ki tarbiyat ke liye istaimaal kiya ja sakta hai. Aap is data par zahaanat laane ke liye algorithm chala sakte hain. Ye aap ki gari ko ye faisla karne ke qabil banaega ke kab murna hai, konsa rasta akhtyar karna hai, kab speed kam karni hai aur kab speed tez karni hai.
- Pesh goi karne ke liye machine learning(Machine learning for making predictions) - Agar aap ke pass finance company ke len den ka data hai aur mustaqbil ke rujhan ka taiyun karne ke liye model banaane ki zaroorat hai to machine learning algorithm behtreen shart hai. Ye zer nigraani seekhne ke namoone ke tahat aata hai. Ise supervised kaha jata hai kiyoon ke aap ke pas pehle se data mojood hai jis ki bunyad par aap apni machino ko tarbiyat de sakte hain. Misaal ke tor par, dhoka dahi ka pata lagaane waale model ko dhoka dahi ki khareedari ke taareekhi record ka instaimaal karte hue tarbiyat dee ja sakti hai.
- Pattern ya namoonon ya khaakon ki daryaft ke liye machine learning (Machine learning for pattern discovery ) - Agar aap ke pas ese parameters nahi hain jin ki bunyad par aap peshan goiyan kar sakte hain, to aap ko data set ke andar chupay hue pattern talaash karne ki zaroorat hai taake maane khaiz pesh goiyan ki ja saken. Ye ghair nigraani shuda model ke siwa kuch nahi hai, kiyoon ke aap ke pas group bandi ke liye pehle se te shuda label nahi hai. Pattern ki daryaft ke liye istaimaal hone wala sab se aam algorithm clustering hai. Hum kehte hain ke aap ek telephone company main kaam kar rahe hain aur aap ko kisi ilaaqe main towers laga kar network qaim karne ki zaroorat hai. Us ke baad clustering techniq ka istaimaal un tower maqamat ko talash karne ke liye kar sakte hain jo is baat ko yaqeeni banaega ke tamam sarifeen signal ki zyada se zyada taqat hasil karen.