2ar1072024, 16:29 Honour_Text_lassfcation_for_Sentiment ip inst nite setkst-learn pandss FE Aequirenent a1ready satisfied: alte in /usr/local/tsb/eythons.20/dist-packages (3.8.1) Requirenent ainesdy satisfied: scikst-leaen in /usr/loea)/Lie/pyenend1@/cist-paccages (1.5.2) Requirenent already Satisfied: pandas in /ust/Local/1ib/python3-10/aist-paceages (2.2.2), Recvirenont already satisfied: Elick in /usr/Local/1ib/pythons.18/¢i5t-packages (éron nish) (81.7) cs (Fram ith) (2.4.2) Requirement aineady satisfied: regox>=2023.8.9 in /usr/local/16/python®.o/dese-pockages (fron nitk) (2026.9.11) Requirement already satisfied: tqdn In /ust/loeal/lib/oython).28/dist-packages (Fron Alli) (6-665) nt already satiafied: munpyo-1-29.5 in /asr/Toeal/Iib/ python) 28/alst=packages (from acikit-leare) (2.26.4) Recuirenent already Satisfied: Scipy>=1.6.8 in /ase/locai/ib/pythons.i0/dist-paceages (fron scikit-learn) (2-13-1) ‘rrenepoolcel>-3. 1.0 in fusr/loeal/1sb/pythons.s0/dise-packagee (Aram seSktt-loaea) (2-5-0) 2.8.2 tn /usr/1oCal/11b/pytnon3.10/eist-pactages (#ron pandas) (2-8-2) 8-1 in /vsr/locai/Mib/pytnon3.18/eist-paceages (From paras) (2624.2) Requirement already satisfied: €24ata>=2022.7 9 fuse/local/1sb/aython3-40/dst-packages (Fron pandas) (2024-2) Recuirenent already satisfied: shx3-1.3 sn /usr/Toesl/Isb/python}.28/akst-packages (fron python-dateutsl>=2-2.2-2p9r838) (1.16.9) Requirement already satisfied: Joblsb in /sr/loeal/1i0/pyenon3. 10/4138 pack Requirenent ainendy satisfied Requirement already satisfied Requirement already satisfied tevti anport re fnport pandas a5 pe Smport nite ‘tran skearnfeature.extraction. text ingort T#idfVectortzer rom skLearn model selection sngort train test split rom sklearncreighbors import tNeightorsClassi tier ‘rom sklearnonetries Inport classification-report, accuracy. score rom sklearn-preprocessing import LabelEncoder # bamiond nite esta rithedounlondt'stoprords') sith download ‘purke") Sample Tustter Data (replace with actual dataset or use Taitter API) asta = ( Sweet's [ “r love this product! e's anazing. #hagoy", vinis is the worst experience £ have ever had! #angry", "I'm not sure how I feel sbout this. I's okay, T guess. treutral”, ‘his 44 fantastic! Great Job. #hIPEY"y Feeling good about this. Wat a great day. Ynapey”, "This product is terrible. Tt broke in two aye. angry", 1's okay, tot the best but not the worst ether. mneutral”, L ercinent positive’, "Negstive’, "Nevtral’, ‘Negative’, positive’, "Positive’, “Negstive’, “Neutral 2 af = pa.batarrane(aata) ‘SE [nith dnt} oowoading package stopords to /rot/nith date {aicaata) " Untpping corgore/stopmords. 20. {nuthcdsta] Bounoasing package purkt to /rot/nite (rithcdate) — Urzipping Eokentzers/purkt. 2. 1 step 1: Prepracess the Text date det prepracess_text(text) I convert to lowercase ‘ext = text Lowerd) ext = re sub(rnttp\se baa\seltetps\se", "4 text, flagsere.MULTILINE) 4 fnove hashtags and mentions text = ressub(e" iar [\#","", text) ext = re-Sub(r Wy text) text = recsub(re de", °°) text) text = ressub(e'\se, * *5 text). steip() + sonly preprocessing -ntps:ifolab research google. comvdrveingkMUGO2NgDBW8ASoKh2PCGwanhsYUXfprniMede=true 18.2ar1072024, 16:29 Honour_Text_t AF ("Cleaned twnet*] = of[ 1466" ].apply(prepracess sex) 4 seep 2: Convert Text bata t9 TF-IDE vectors {lar yectorszer = Teidflectorizer(nax_featuress500, stop sordsestopuondsworda(‘enghsh')) Xie Clue vectortzer.._transforn(ae[ Cheaned Tweet” })-tourray() 1 encode sentinant Labels into numerical format abel encoder = Labelencoder() y's label encoder. (it_traneforn(4f('Sentinent’]) Onlegative, Aotevteal, 2ePositive 4 seep 3: Split she oataset into Training and Testing Oats K.train, Xtest, yotrain, y.test = train test split, yp test_size-e.2, random staton42) ken_elassifser = KeighborsChassifier(n_nesghtorse3) en_elassiier.#2U0\ train, y_traie) + Kctgnborsctassiier rveighborsclarssfter(n_neighborse3) 1 seep 5: fake Poestetions on Test oats Yupred = kan_clazeifien.proekce (seat) print(*\nclaselfiestion Report:\n") 1 cet unique Labels on both y_test andy. pred unique labels = Sorced(List(set(y_test) | Set(y_oree))) fF ralter carget nanes based on these unique 2abeis tonget_naree = (Jabel_encoderclaeses_[1) for £ 39 unique_1nbels) F Generate repore with filtered target names and labels print(classifieation.roport(y test, y.pred, target nanesetargst_nanes, Isbels-unique_120e15)) accuracy + scearacy_seore(y.test, y_ared) prine(eaccursey: {accuracy * 300:-24)) asst fication Report: negative 3.08 2.08.0 2 Posstive teehee Tbe i macro og 3.08 2.82.88 2 weighted ve 881.88 1.08 2 Accuracy: 160.005 + branple to predict sentinert of new tucets new_ewoets = ["I hate aaiting for #0 long! This Ss frustrating.*, “Wos, this Se fantasticl Love $¢.°] rnew_elosned tweots = [praprocess_text(twest) for tweet in neu. tweets] new_tfiet = tflde vectorizer-transtorn(nen cleaned tweets) -tosrray() rewcpredictions ~ kan elassifler predict (new tf14F) rescsentinente ~ label encoder. inverse. transforn(nex_predtctSons) brint(*\nsentinant Pradictions for New Tweets:*) for tweet, sertinent in zip(rew tweets, ew sentiments): prine(e"Tweet: {eweet) \nPredicted Sentinent: (sentanent)\0") a Santinant Predictions for New Tweets Predicted Sentiment: Neutral -ntps:ifolab research google. comvdrveingkMUGO2NgDBW8ASoKh2PCGwanhsYUXfprniMede=true nalysis_using_KNN_Note 2682ar1072024, 16:29 Honour_Text_classfcation for_Sentime mnalysis_using_KNN._Note_Use_twiter_datajpynb -Colab hips oolab research google. comidrverInqkMUGO2NgD8t#ASoKN2PCGwanthsYUXipiniMode=tre 38.