Node 08. 모델 학습 기술

8-1. 들어가며

---

학습 목표

• 딥러닝 모델 학습의 개념 및 기술 파악
• 과소적합 및 과대적합에 대한 설명
• IMDB 데이터셋 이용 : 긍정/부정 분류 딥러닝 모델 생성

학습 내용

• 모델 학습 기술

  • 학습 단위
  • 데이터 스케줄링
  • 학습률과 에폭
  • 은닉층과 뉴런수
  • 활성화 함수
  • 가중치 초기화
  • 옵티마이저

• 과소적합과 과대적합

  • 과소적합
  • 과대적합
  • 과대적합과 과소적합 방지 방법

• IMDB 딥러닝 모델 예제

  • 데이터 로드 및 전처리
  • 모델 구성
  • 모델 컴파일 및 학습

8-2. 모델 학습 기술

---

모델 학습 기술 정리

학습 단위

• 전체 배치(Full Batch)

  • 전통적 통계 기법
  • 모든 데이터를 알고리즘에 한꺼번에 넣어서 분석

• Stochastic

  • 전체 데이터 중 랜덤 추출해 알고리즘에 넣는 방법

• Mini Batch(미니 배치)

  • 전체 데이터 중 일부를 모델에 사용
  • 미니 배치 크기 : 한번 학습 시의 크기 결정
  • 배치 크기 작을수록 ➡️ 학습 시간 많이 소요(trade-off)
    

Q. 전체 배치 방식이 있음에도, 미니 배치를 사용하는 이유

A. 딥러닝의 경우, 대용량의 데이터를 사용하기 때문에 모든 데이터를 한번에 로드해서 학습하게되면 리소스 낭비가 크기 때문

Data Scaling

• 데이터 스케일이 다르면 -> 딥러닝 모델 학습 시 동작이 잘 안됨

• 모든 특성의 범위 또는 분포는 같게 해주는 것이 바람직

• 표준화(Standardization)

  • 주로 사용
  • 특성 평균을 0, 분산을 1로 스케일링
  • 특성들의 정규분포를 만듦

• 정규화(Normalization)

  • 특성들을 -> 특정 범위로 스케일링
  • 가장 작은 값을 0, 가장 큰 값을 1로 변환해 0 ~ 1 범위를 갖게 하는 방식

Q. 표준화(Standardization) vs 정규화(Normalization)

A. 표준화는 평균을 0으로 만드는 것이고, 분산을 1로 만들지만 정규화는 최솟값을 0, 최댓값을 1로 만드는 것에 차이가 있음

LearningRate & Epochs(학습률과 에폭)

• 학습률 : 가중치 조정 단위, 오차 최소화를 위해 간격을 결정해야 함
• 학습률이 너무 클 경우
  • 최솟값을 오버슈트(Overshoot)
• 학습률이 너무 작을 경우
  • 최솟값에 이르기 전 학습 중단 우려

• 에폭 : 1번의 학습 과정을 몇 번 수행할 것인지 지정
  • 너무 작게 하거나 너무 크게 하면 -> 과소 적합 / 과대 적합 발생
  • 진행을 여러번 해서 최적의 에폭을 찾아야 하는 것

• 학습률과 에폭의 관계
  • 학습률이 적을수록 ➡️ 에폭이 커야함(배치 크기도 고려해야 함)
  • 학습 속도와 연산의 지속 반복 ➡️ cost/loss 변화 살필 것

Q. 손실함수의 최저값에 도달하기 위해, 학습률이 작을 수록 일반적으로 에폭수는 어떻게 달라지는지?

A. 에폭수는 많아짐(업데이트 횟수가 더 많이 필요할 것이기 때문)

Hidden Layer & Neurons(은닉층과 뉴런수)

• 모델 구성 시 고려할 점
  • 은닉층의 구성 방법, 얼마나 쌓을지에 대한 내용, 뉴런수 구성 등
• 이 부분들에 따라 모델 크기 상이
• 학습 및 최적화 정도에도 차이 발생
  

Activation Function(활성화 함수)

• weight(가중치), bias(바이어스) 조정
• 활성화 함수 선택에 따라 ➡️ 모델 학습 정도에 영향이 큼
• 시그모이드(Sigmoid) 계열, ReLU 계열
  

Q. 두 계열간의 차이는?

A. 선형과 비선형

• 시그모이드 계열 : 결과값이 [0,1] or [-1,1] 사이 값으로 나옴
• 렐루 계열 : (예외 제외 시) 0을 중심으로 양수는 그대로, 음수는 0 또는 0과 가까운 수로 나옴

Weight Intialization(가중치 초기화)

• 초기에 가중치 값을 어떻게 할 것인지도 중요
  • 초기값 0 -> 모든 노드들의 값이 같아 여러 노드를 사용하는 의미를 찾기 어려워짐
  • 동일값을 갖지 않도록 무작위 설정 필요
• 가중치 초기화를 위해 확률 분포 사용 & 정규 분포 생성
  • 가중치 값이 일부 값으로 치우침(편향)
  • 활성화 함수를 통과한 값도 치우치고 ➡️ 표현 가능한 신경망 수가 적어짐
• 가중치 초기화는 활성화값이 고루 분포되어야 함!

• 적절한 가중치 초기화 방법

  • Xavier(Glorot)
    • 은닉층 노드수가 n -> 표준 편차가 $1 \over \sqrt{n}$인 분포를 가짐
    • 더 많은 가중치에 역전파 전달 가능
    • 많은 문제 표현 가능
    • 활성화 함수가 선형일 경우 적합
  • He 초기화
    • 표준 편차가 $\sqrt{2 \over n}$인 분포
    • 활성화값이 균일하게 분포됨
    • 활성화 함수가 비선형일 경우 적합

  Q. 적절한 가중치 초기값을 정해주는 것의 효과는?

  A. 표현 가능한 신경망 수가 많아짐, 더 많은 가중치에 역전파 전달 가능, 비교적 많은 문제 표현 가능

Optimizer(옵티마이저)

• 모델이 실제 결과와 예측 결과 차이를 최소화함
  • SGD
    • 확률적 경사 하강법
    • 최소점 찾는 과정에서 멈춤, 속도 느림
    • 모멘텀, 네스테로프 모멘텀, 학습률 감소 기법 지원
  • RMSprop
    • 가장 최근 그래디언트만 반영해 전력 최적점 수렴
    • 지수적 감소를 이용
  • Adagrad
    • 그래디언트 벡터값 스케일링
    • 단위로 인한 경사 하강 오류 수정
    • 기울기 전체 내역 고려
  • Adadelta
    • 너무 빠르게 기울기를 0으로 만드는 Adagrad의 단점 개선
    • 그래디언트 업데이트 이동창 기반으로 학습률 조절
  • Adam
    • 모멘텀 최적화 + RMSprop 병합

Q. 옵티마이저 역할 및 목적

A. 손실함수 감소를 위함(실제 및 예측 값 간의 차이 감소) -> 가중치 업데이트 방식을 결정

8-3. 과소적합과 과대적합

---

UnderFitting

• 학습 데이터가 충분히 학습되지 않아 성능이 아주 좋지 못한 경우를 말함
• 모델의 지나친 단순화로 인해 발생 ➡️ 학습 데이터를 충분히 수집 or 더 복잡한 모델 선택
• epochs를 늘려서 충분히 학습할 수 있게 해야함

OverFitting

• 모델이 학습 데이터에 과하게 맞춰진 경우
• 새로운 데이터에는 성능 저하 발생
• 학습 데이터가 매우 적음, 모델의 지나친 복잡도, 학습 횟수 과도의 경우에 발생
• 다양한 학습 데이터 수집&학습, 파라미터가 적은 모델의 선택, 학습 데이터 특성 수 줄이기 등 ➡️ 모델 단순화가 필요
  

8-4. IMDB 딥러닝 모델 예제

---

IMDB 데이터셋

• 영화 사이트 IMDB 리뷰 데이터 모음
• 텍스트 분류, 감성 분류에 자주 사용
• 리뷰 텍스트 및 레이블로 구성
  • 레이블 : 리뷰가 긍정이면 1, 부정이면 0

데이터 로드 및 전처리

• imdb.load_data() 함수로 케라스 내부에서 다운로드 가능
  • num_words : 10,000개로 제한하여 로드(단어 인덱스는 0 ~ 9999)
• train_data 0번째 출력 : 단어 인덱스 출력됨
• train_label 0번째를 출력 : 긍정 출력(1)

from keras.datasets import imdb
import numpy as np

(train_data, train_labels), (test_data, test_labels) = imdb.load_data(num_words=10000)

print(train_data[0])
print(train_labels[0])

Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb.npz
17465344/17464789 [==============================] - 0s 0us/step
17473536/17464789 [==============================] - 0s 0us/step
[1, 14, 22, 16, 43, 530, 973, 1622, 1385, 65, 458, 4468, 66, 3941, 4, 173, 36, 256, 5, 25, 100, 43, 838, 112, 50, 670, 2, 9, 35, 480, 284, 5, 150, 4, 172, 112, 167, 2, 336, 385, 39, 4, 172, 4536, 1111, 17, 546, 38, 13, 447, 4, 192, 50, 16, 6, 147, 2025, 19, 14, 22, 4, 1920, 4613, 469, 4, 22, 71, 87, 12, 16, 43, 530, 38, 76, 15, 13, 1247, 4, 22, 17, 515, 17, 12, 16, 626, 18, 2, 5, 62, 386, 12, 8, 316, 8, 106, 5, 4, 2223, 5244, 16, 480, 66, 3785, 33, 4, 130, 12, 16, 38, 619, 5, 25, 124, 51, 36, 135, 48, 25, 1415, 33, 6, 22, 12, 215, 28, 77, 52, 5, 14, 407, 16, 82, 2, 8, 4, 107, 117, 5952, 15, 256, 4, 2, 7, 3766, 5, 723, 36, 71, 43, 530, 476, 26, 400, 317, 46, 7, 4, 2, 1029, 13, 104, 88, 4, 381, 15, 297, 98, 32, 2071, 56, 26, 141, 6, 194, 7486, 18, 4, 226, 22, 21, 134, 476, 26, 480, 5, 144, 30, 5535, 18, 51, 36, 28, 224, 92, 25, 104, 4, 226, 65, 16, 38, 1334, 88, 12, 16, 283, 5, 16, 4472, 113, 103, 32, 15, 16, 5345, 19, 178, 32]
1

• 실제 단어 변환 : 단어 인덱스 딕셔너리 이용해야 함
  • imdb.get_word_index()

word_index = imdb.get_word_index()
word_index

Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb_word_index.json
1646592/1641221 [==============================] - 0s 0us/step
1654784/1641221 [==============================] - 0s 0us/step

{'fawn': 34701,
 'tsukino': 52006,
 'nunnery': 52007,
 'sonja': 16816,
 'vani': 63951,
 'woods': 1408,
 'spiders': 16115,
 'hanging': 2345,
 'woody': 2289,
 'trawling': 52008,
 "hold's": 52009,
 'comically': 11307,
 'localized': 40830,
 'disobeying': 30568,
 "'royale": 52010,
 "harpo's": 40831,
 'canet': 52011,
 'aileen': 19313,
 'acurately': 52012,
 "diplomat's": 52013,
 'rickman': 25242,
 'arranged': 6746,
 'rumbustious': 52014,
 'familiarness': 52015,
 "spider'": 52016,
 'hahahah': 68804,
 "wood'": 52017,
 'transvestism': 40833,
 "hangin'": 34702,
 'bringing': 2338,
 'seamier': 40834,
 'wooded': 34703,
 'bravora': 52018,
 'grueling': 16817,
 'wooden': 1636,
 'wednesday': 16818,
 "'prix": 52019,
 'altagracia': 34704,
 'circuitry': 52020,
 'crotch': 11585,
 'busybody': 57766,
 "tart'n'tangy": 52021,
 'burgade': 14129,
 'thrace': 52023,
 "tom's": 11038,
 'snuggles': 52025,
 'francesco': 29114,
 'complainers': 52027,
 'templarios': 52125,
 '272': 40835,
 '273': 52028,
 'zaniacs': 52130,
 '275': 34706,
 'consenting': 27631,
 'snuggled': 40836,
 'inanimate': 15492,
 'uality': 52030,
 'bronte': 11926,
 'errors': 4010,
 'dialogs': 3230,
 "yomada's": 52031,
 "madman's": 34707,
 'dialoge': 30585,
 'usenet': 52033,
 'videodrome': 40837,
 "kid'": 26338,
 'pawed': 52034,
 "'girlfriend'": 30569,
 "'pleasure": 52035,
 "'reloaded'": 52036,
 "kazakos'": 40839,
 'rocque': 52037,
 'mailings': 52038,
 'brainwashed': 11927,
 'mcanally': 16819,
 "tom''": 52039,
 'kurupt': 25243,
 'affiliated': 21905,
 'babaganoosh': 52040,
 "noe's": 40840,
 'quart': 40841,
 'kids': 359,
 'uplifting': 5034,
 'controversy': 7093,
 'kida': 21906,
 'kidd': 23379,
 "error'": 52041,
 'neurologist': 52042,
 'spotty': 18510,
 'cobblers': 30570,
 'projection': 9878,
 'fastforwarding': 40842,
 'sters': 52043,
 "eggar's": 52044,
 'etherything': 52045,
 'gateshead': 40843,
 'airball': 34708,
 'unsinkable': 25244,
 'stern': 7180,
 "cervi's": 52046,
 'dnd': 40844,
 'dna': 11586,
 'insecurity': 20598,
 "'reboot'": 52047,
 'trelkovsky': 11037,
 'jaekel': 52048,
 'sidebars': 52049,
 "sforza's": 52050,
 'distortions': 17633,
 'mutinies': 52051,
 'sermons': 30602,
 '7ft': 40846,
 'boobage': 52052,
 "o'bannon's": 52053,
 'populations': 23380,
 'chulak': 52054,
 'mesmerize': 27633,
 'quinnell': 52055,
 'yahoo': 10307,
 'meteorologist': 52057,
 'beswick': 42577,
 'boorman': 15493,
 'voicework': 40847,
 "ster'": 52058,
 'blustering': 22922,
 'hj': 52059,
 'intake': 27634,
 'morally': 5621,
 'jumbling': 40849,
 'bowersock': 52060,
 "'porky's'": 52061,
 'gershon': 16821,
 'ludicrosity': 40850,
 'coprophilia': 52062,
 'expressively': 40851,
 "india's": 19500,
 "post's": 34710,
 'wana': 52063,
 'wang': 5283,
 'wand': 30571,
 'wane': 25245,
 'edgeways': 52321,
 'titanium': 34711,
 'pinta': 40852,
 'want': 178,
 'pinto': 30572,
 'whoopdedoodles': 52065,
 'tchaikovsky': 21908,
 'travel': 2103,
 "'victory'": 52066,
 'copious': 11928,
 'gouge': 22433,
 "chapters'": 52067,
 'barbra': 6702,
 'uselessness': 30573,
 "wan'": 52068,
 'assimilated': 27635,
 'petiot': 16116,
 'most\x85and': 52069,
 'dinosaurs': 3930,
 'wrong': 352,
 'seda': 52070,
 'stollen': 52071,
 'sentencing': 34712,
 'ouroboros': 40853,
 'assimilates': 40854,
 'colorfully': 40855,
 'glenne': 27636,
 'dongen': 52072,
 'subplots': 4760,
 'kiloton': 52073,
 'chandon': 23381,
 "effect'": 34713,
 'snugly': 27637,
 'kuei': 40856,
 'welcomed': 9092,
 'dishonor': 30071,
 'concurrence': 52075,
 'stoicism': 23382,
 "guys'": 14896,
 "beroemd'": 52077,
 'butcher': 6703,
 "melfi's": 40857,
 'aargh': 30623,
 'playhouse': 20599,
 'wickedly': 11308,
 'fit': 1180,
 'labratory': 52078,
 'lifeline': 40859,
 'screaming': 1927,
 'fix': 4287,
 'cineliterate': 52079,
 'fic': 52080,
 'fia': 52081,
 'fig': 34714,
 'fmvs': 52082,
 'fie': 52083,
 'reentered': 52084,
 'fin': 30574,
 'doctresses': 52085,
 'fil': 52086,
 'zucker': 12606,
 'ached': 31931,
 'counsil': 52088,
 'paterfamilias': 52089,
 'songwriter': 13885,
 'shivam': 34715,
 'hurting': 9654,
 'effects': 299,
 'slauther': 52090,
 "'flame'": 52091,
 'sommerset': 52092,
 'interwhined': 52093,
 'whacking': 27638,
 'bartok': 52094,
 'barton': 8775,
 'frewer': 21909,
 "fi'": 52095,
 'ingrid': 6192,
 'stribor': 30575,
 'approporiately': 52096,
 'wobblyhand': 52097,
 'tantalisingly': 52098,
 'ankylosaurus': 52099,
 'parasites': 17634,
 'childen': 52100,
 "jenkins'": 52101,
 'metafiction': 52102,
 'golem': 17635,
 'indiscretion': 40860,
 "reeves'": 23383,
 "inamorata's": 57781,
 'brittannica': 52104,
 'adapt': 7916,
 "russo's": 30576,
 'guitarists': 48246,
 'abbott': 10553,
 'abbots': 40861,
 'lanisha': 17649,
 'magickal': 40863,
 'mattter': 52105,
 "'willy": 52106,
 'pumpkins': 34716,
 'stuntpeople': 52107,
 'estimate': 30577,
 'ugghhh': 40864,
 'gameplay': 11309,
 "wern't": 52108,
 "n'sync": 40865,
 'sickeningly': 16117,
 'chiara': 40866,
 'disturbed': 4011,
 'portmanteau': 40867,
 'ineffectively': 52109,
 "duchonvey's": 82143,
 "nasty'": 37519,
 'purpose': 1285,
 'lazers': 52112,
 'lightened': 28105,
 'kaliganj': 52113,
 'popularism': 52114,
 "damme's": 18511,
 'stylistics': 30578,
 'mindgaming': 52115,
 'spoilerish': 46449,
 "'corny'": 52117,
 'boerner': 34718,
 'olds': 6792,
 'bakelite': 52118,
 'renovated': 27639,
 'forrester': 27640,
 "lumiere's": 52119,
 'gaskets': 52024,
 'needed': 884,
 'smight': 34719,
 'master': 1297,
 "edie's": 25905,
 'seeber': 40868,
 'hiya': 52120,
 'fuzziness': 52121,
 'genesis': 14897,
 'rewards': 12607,
 'enthrall': 30579,
 "'about": 40869,
 "recollection's": 52122,
 'mutilated': 11039,
 'fatherlands': 52123,
 "fischer's": 52124,
 'positively': 5399,
 '270': 34705,
 'ahmed': 34720,
 'zatoichi': 9836,
 'bannister': 13886,
 'anniversaries': 52127,
 "helm's": 30580,
 "'work'": 52128,
 'exclaimed': 34721,
 "'unfunny'": 52129,
 '274': 52029,
 'feeling': 544,
 "wanda's": 52131,
 'dolan': 33266,
 '278': 52133,
 'peacoat': 52134,
 'brawny': 40870,
 'mishra': 40871,
 'worlders': 40872,
 'protags': 52135,
 'skullcap': 52136,
 'dastagir': 57596,
 'affairs': 5622,
 'wholesome': 7799,
 'hymen': 52137,
 'paramedics': 25246,
 'unpersons': 52138,
 'heavyarms': 52139,
 'affaire': 52140,
 'coulisses': 52141,
 'hymer': 40873,
 'kremlin': 52142,
 'shipments': 30581,
 'pixilated': 52143,
 "'00s": 30582,
 'diminishing': 18512,
 'cinematic': 1357,
 'resonates': 14898,
 'simplify': 40874,
 "nature'": 40875,
 'temptresses': 40876,
 'reverence': 16822,
 'resonated': 19502,
 'dailey': 34722,
 '2\x85': 52144,
 'treize': 27641,
 'majo': 52145,
 'kiya': 21910,
 'woolnough': 52146,
 'thanatos': 39797,
 'sandoval': 35731,
 'dorama': 40879,
 "o'shaughnessy": 52147,
 'tech': 4988,
 'fugitives': 32018,
 'teck': 30583,
 "'e'": 76125,
 'doesn’t': 40881,
 'purged': 52149,
 'saying': 657,
 "martians'": 41095,
 'norliss': 23418,
 'dickey': 27642,
 'dicker': 52152,
 "'sependipity": 52153,
 'padded': 8422,
 'ordell': 57792,
 "sturges'": 40882,
 'independentcritics': 52154,
 'tempted': 5745,
 "atkinson's": 34724,
 'hounded': 25247,
 'apace': 52155,
 'clicked': 15494,
 "'humor'": 30584,
 "martino's": 17177,
 "'supporting": 52156,
 'warmongering': 52032,
 "zemeckis's": 34725,
 'lube': 21911,
 'shocky': 52157,
 'plate': 7476,
 'plata': 40883,
 'sturgess': 40884,
 "nerds'": 40885,
 'plato': 20600,
 'plath': 34726,
 'platt': 40886,
 'mcnab': 52159,
 'clumsiness': 27643,
 'altogether': 3899,
 'massacring': 42584,
 'bicenntinial': 52160,
 'skaal': 40887,
 'droning': 14360,
 'lds': 8776,
 'jaguar': 21912,
 "cale's": 34727,
 'nicely': 1777,
 'mummy': 4588,
 "lot's": 18513,
 'patch': 10086,
 'kerkhof': 50202,
 "leader's": 52161,
 "'movie": 27644,
 'uncomfirmed': 52162,
 'heirloom': 40888,
 'wrangle': 47360,
 'emotion\x85': 52163,
 "'stargate'": 52164,
 'pinoy': 40889,
 'conchatta': 40890,
 'broeke': 41128,
 'advisedly': 40891,
 "barker's": 17636,
 'descours': 52166,
 'lots': 772,
 'lotr': 9259,
 'irs': 9879,
 'lott': 52167,
 'xvi': 40892,
 'irk': 34728,
 'irl': 52168,
 'ira': 6887,
 'belzer': 21913,
 'irc': 52169,
 'ire': 27645,
 'requisites': 40893,
 'discipline': 7693,
 'lyoko': 52961,
 'extend': 11310,
 'nature': 873,
 "'dickie'": 52170,
 'optimist': 40894,
 'lapping': 30586,
 'superficial': 3900,
 'vestment': 52171,
 'extent': 2823,
 'tendons': 52172,
 "heller's": 52173,
 'quagmires': 52174,
 'miyako': 52175,
 'moocow': 20601,
 "coles'": 52176,
 'lookit': 40895,
 'ravenously': 52177,
 'levitating': 40896,
 'perfunctorily': 52178,
 'lookin': 30587,
 "lot'": 40898,
 'lookie': 52179,
 'fearlessly': 34870,
 'libyan': 52181,
 'fondles': 40899,
 'gopher': 35714,
 'wearying': 40901,
 "nz's": 52182,
 'minuses': 27646,
 'puposelessly': 52183,
 'shandling': 52184,
 'decapitates': 31268,
 'humming': 11929,
 "'nother": 40902,
 'smackdown': 21914,
 'underdone': 30588,
 'frf': 40903,
 'triviality': 52185,
 'fro': 25248,
 'bothers': 8777,
 "'kensington": 52186,
 'much': 73,
 'muco': 34730,
 'wiseguy': 22615,
 "richie's": 27648,
 'tonino': 40904,
 'unleavened': 52187,
 'fry': 11587,
 "'tv'": 40905,
 'toning': 40906,
 'obese': 14361,
 'sensationalized': 30589,
 'spiv': 40907,
 'spit': 6259,
 'arkin': 7364,
 'charleton': 21915,
 'jeon': 16823,
 'boardroom': 21916,
 'doubts': 4989,
 'spin': 3084,
 'hepo': 53083,
 'wildcat': 27649,
 'venoms': 10584,
 'misconstrues': 52191,
 'mesmerising': 18514,
 'misconstrued': 40908,
 'rescinds': 52192,
 'prostrate': 52193,
 'majid': 40909,
 'climbed': 16479,
 'canoeing': 34731,
 'majin': 52195,
 'animie': 57804,
 'sylke': 40910,
 'conditioned': 14899,
 'waddell': 40911,
 '3\x85': 52196,
 'hyperdrive': 41188,
 'conditioner': 34732,
 'bricklayer': 53153,
 'hong': 2576,
 'memoriam': 52198,
 'inventively': 30592,
 "levant's": 25249,
 'portobello': 20638,
 'remand': 52200,
 'mummified': 19504,
 'honk': 27650,
 'spews': 19505,
 'visitations': 40912,
 'mummifies': 52201,
 'cavanaugh': 25250,
 'zeon': 23385,
 "jungle's": 40913,
 'viertel': 34733,
 'frenchmen': 27651,
 'torpedoes': 52202,
 'schlessinger': 52203,
 'torpedoed': 34734,
 'blister': 69876,
 'cinefest': 52204,
 'furlough': 34735,
 'mainsequence': 52205,
 'mentors': 40914,
 'academic': 9094,
 'stillness': 20602,
 'academia': 40915,
 'lonelier': 52206,
 'nibby': 52207,
 "losers'": 52208,
 'cineastes': 40916,
 'corporate': 4449,
 'massaging': 40917,
 'bellow': 30593,
 'absurdities': 19506,
 'expetations': 53241,
 'nyfiken': 40918,
 'mehras': 75638,
 'lasse': 52209,
 'visability': 52210,
 'militarily': 33946,
 "elder'": 52211,
 'gainsbourg': 19023,
 'hah': 20603,
 'hai': 13420,
 'haj': 34736,
 'hak': 25251,
 'hal': 4311,
 'ham': 4892,
 'duffer': 53259,
 'haa': 52213,
 'had': 66,
 'advancement': 11930,
 'hag': 16825,
 "hand'": 25252,
 'hay': 13421,
 'mcnamara': 20604,
 "mozart's": 52214,
 'duffel': 30731,
 'haq': 30594,
 'har': 13887,
 'has': 44,
 'hat': 2401,
 'hav': 40919,
 'haw': 30595,
 'figtings': 52215,
 'elders': 15495,
 'underpanted': 52216,
 'pninson': 52217,
 'unequivocally': 27652,
 "barbara's": 23673,
 "bello'": 52219,
 'indicative': 12997,
 'yawnfest': 40920,
 'hexploitation': 52220,
 "loder's": 52221,
 'sleuthing': 27653,
 "justin's": 32622,
 "'ball": 52222,
 "'summer": 52223,
 "'demons'": 34935,
 "mormon's": 52225,
 "laughton's": 34737,
 'debell': 52226,
 'shipyard': 39724,
 'unabashedly': 30597,
 'disks': 40401,
 'crowd': 2290,
 'crowe': 10087,
 "vancouver's": 56434,
 'mosques': 34738,
 'crown': 6627,
 'culpas': 52227,
 'crows': 27654,
 'surrell': 53344,
 'flowless': 52229,
 'sheirk': 52230,
 "'three": 40923,
 "peterson'": 52231,
 'ooverall': 52232,
 'perchance': 40924,
 'bottom': 1321,
 'chabert': 53363,
 'sneha': 52233,
 'inhuman': 13888,
 'ichii': 52234,
 'ursla': 52235,
 'completly': 30598,
 'moviedom': 40925,
 'raddick': 52236,
 'brundage': 51995,
 'brigades': 40926,
 'starring': 1181,
 "'goal'": 52237,
 'caskets': 52238,
 'willcock': 52239,
 "threesome's": 52240,
 "mosque'": 52241,
 "cover's": 52242,
 'spaceships': 17637,
 'anomalous': 40927,
 'ptsd': 27655,
 'shirdan': 52243,
 'obscenity': 21962,
 'lemmings': 30599,
 'duccio': 30600,
 "levene's": 52244,
 "'gorby'": 52245,
 "teenager's": 25255,
 'marshall': 5340,
 'honeymoon': 9095,
 'shoots': 3231,
 'despised': 12258,
 'okabasho': 52246,
 'fabric': 8289,
 'cannavale': 18515,
 'raped': 3537,
 "tutt's": 52247,
 'grasping': 17638,
 'despises': 18516,
 "thief's": 40928,
 'rapes': 8926,
 'raper': 52248,
 "eyre'": 27656,
 'walchek': 52249,
 "elmo's": 23386,
 'perfumes': 40929,
 'spurting': 21918,
 "exposition'\x85": 52250,
 'denoting': 52251,
 'thesaurus': 34740,
 "shoot'": 40930,
 'bonejack': 49759,
 'simpsonian': 52253,
 'hebetude': 30601,
 "hallow's": 34741,
 'desperation\x85': 52254,
 'incinerator': 34742,
 'congratulations': 10308,
 'humbled': 52255,
 "else's": 5924,
 'trelkovski': 40845,
 "rape'": 52256,
 "'chapters'": 59386,
 '1600s': 52257,
 'martian': 7253,
 'nicest': 25256,
 'eyred': 52259,
 'passenger': 9457,
 'disgrace': 6041,
 'moderne': 52260,
 'barrymore': 5120,
 'yankovich': 52261,
 'moderns': 40931,
 'studliest': 52262,
 'bedsheet': 52263,
 'decapitation': 14900,
 'slurring': 52264,
 "'nunsploitation'": 52265,
 "'character'": 34743,
 'cambodia': 9880,
 'rebelious': 52266,
 'pasadena': 27657,
 'crowne': 40932,
 "'bedchamber": 52267,
 'conjectural': 52268,
 'appologize': 52269,
 'halfassing': 52270,
 'paycheque': 57816,
 'palms': 20606,
 "'islands": 52271,
 'hawked': 40933,
 'palme': 21919,
 'conservatively': 40934,
 'larp': 64007,
 'palma': 5558,
 'smelling': 21920,
 'aragorn': 12998,
 'hawker': 52272,
 'hawkes': 52273,
 'explosions': 3975,
 'loren': 8059,
 "pyle's": 52274,
 'shootout': 6704,
 "mike's": 18517,
 "driscoll's": 52275,
 'cogsworth': 40935,
 "britian's": 52276,
 'childs': 34744,
 "portrait's": 52277,
 'chain': 3626,
 'whoever': 2497,
 'puttered': 52278,
 'childe': 52279,
 'maywether': 52280,
 'chair': 3036,
 "rance's": 52281,
 'machu': 34745,
 'ballet': 4517,
 'grapples': 34746,
 'summerize': 76152,
 'freelance': 30603,
 "andrea's": 52283,
 '\x91very': 52284,
 'coolidge': 45879,
 'mache': 18518,
 'balled': 52285,
 'grappled': 40937,
 'macha': 18519,
 'underlining': 21921,
 'macho': 5623,
 'oversight': 19507,
 'machi': 25257,
 'verbally': 11311,
 'tenacious': 21922,
 'windshields': 40938,
 'paychecks': 18557,
 'jerk': 3396,
 "good'": 11931,
 'prancer': 34748,
 'prances': 21923,
 'olympus': 52286,
 'lark': 21924,
 'embark': 10785,
 'gloomy': 7365,
 'jehaan': 52287,
 'turaqui': 52288,
 "child'": 20607,
 'locked': 2894,
 'pranced': 52289,
 'exact': 2588,
 'unattuned': 52290,
 'minute': 783,
 'skewed': 16118,
 'hodgins': 40940,
 'skewer': 34749,
 'think\x85': 52291,
 'rosenstein': 38765,
 'helmit': 52292,
 'wrestlemanias': 34750,
 'hindered': 16826,
 "martha's": 30604,
 'cheree': 52293,
 "pluckin'": 52294,
 'ogles': 40941,
 'heavyweight': 11932,
 'aada': 82190,
 'chopping': 11312,
 'strongboy': 61534,
 'hegemonic': 41342,
 'adorns': 40942,
 'xxth': 41346,
 'nobuhiro': 34751,
 'capitães': 52298,
 'kavogianni': 52299,
 'antwerp': 13422,
 'celebrated': 6538,
 'roarke': 52300,
 'baggins': 40943,
 'cheeseburgers': 31270,
 'matras': 52301,
 "nineties'": 52302,
 "'craig'": 52303,
 'celebrates': 12999,
 'unintentionally': 3383,
 'drafted': 14362,
 'climby': 52304,
 '303': 52305,
 'oldies': 18520,
 'climbs': 9096,
 'honour': 9655,
 'plucking': 34752,
 '305': 30074,
 'address': 5514,
 'menjou': 40944,
 "'freak'": 42592,
 'dwindling': 19508,
 'benson': 9458,
 'white’s': 52307,
 'shamelessness': 40945,
 'impacted': 21925,
 'upatz': 52308,
 'cusack': 3840,
 "flavia's": 37567,
 'effette': 52309,
 'influx': 34753,
 'boooooooo': 52310,
 'dimitrova': 52311,
 'houseman': 13423,
 'bigas': 25259,
 'boylen': 52312,
 'phillipenes': 52313,
 'fakery': 40946,
 "grandpa's": 27658,
 'darnell': 27659,
 'undergone': 19509,
 'handbags': 52315,
 'perished': 21926,
 'pooped': 37778,
 'vigour': 27660,
 'opposed': 3627,
 'etude': 52316,
 "caine's": 11799,
 'doozers': 52317,
 'photojournals': 34754,
 'perishes': 52318,
 'constrains': 34755,
 'migenes': 40948,
 'consoled': 30605,
 'alastair': 16827,
 'wvs': 52319,
 'ooooooh': 52320,
 'approving': 34756,
 'consoles': 40949,
 'disparagement': 52064,
 'futureistic': 52322,
 'rebounding': 52323,
 "'date": 52324,
 'gregoire': 52325,
 'rutherford': 21927,
 'americanised': 34757,
 'novikov': 82196,
 'following': 1042,
 'munroe': 34758,
 "morita'": 52326,
 'christenssen': 52327,
 'oatmeal': 23106,
 'fossey': 25260,
 'livered': 40950,
 'listens': 13000,
 "'marci": 76164,
 "otis's": 52330,
 'thanking': 23387,
 'maude': 16019,
 'extensions': 34759,
 'ameteurish': 52332,
 "commender's": 52333,
 'agricultural': 27661,
 'convincingly': 4518,
 'fueled': 17639,
 'mahattan': 54014,
 "paris's": 40952,
 'vulkan': 52336,
 'stapes': 52337,
 'odysessy': 52338,
 'harmon': 12259,
 'surfing': 4252,
 'halloran': 23494,
 'unbelieveably': 49580,
 "'offed'": 52339,
 'quadrant': 30607,
 'inhabiting': 19510,
 'nebbish': 34760,
 'forebears': 40953,
 'skirmish': 34761,
 'ocassionally': 52340,
 "'resist": 52341,
 'impactful': 21928,
 'spicier': 52342,
 'touristy': 40954,
 "'football'": 52343,
 'webpage': 40955,
 'exurbia': 52345,
 'jucier': 52346,
 'professors': 14901,
 'structuring': 34762,
 'jig': 30608,
 'overlord': 40956,
 'disconnect': 25261,
 'sniffle': 82201,
 'slimeball': 40957,
 'jia': 40958,
 'milked': 16828,
 'banjoes': 40959,
 'jim': 1237,
 'workforces': 52348,
 'jip': 52349,
 'rotweiller': 52350,
 'mundaneness': 34763,
 "'ninja'": 52351,
 "dead'": 11040,
 "cipriani's": 40960,
 'modestly': 20608,
 "professor'": 52352,
 'shacked': 40961,
 'bashful': 34764,
 'sorter': 23388,
 'overpowering': 16120,
 'workmanlike': 18521,
 'henpecked': 27662,
 'sorted': 18522,
 "jōb's": 52354,
 "'always": 52355,
 "'baptists": 34765,
 'dreamcatchers': 52356,
 "'silence'": 52357,
 'hickory': 21929,
 'fun\x97yet': 52358,
 'breakumentary': 52359,
 'didn': 15496,
 'didi': 52360,
 'pealing': 52361,
 'dispite': 40962,
 "italy's": 25262,
 'instability': 21930,
 'quarter': 6539,
 'quartet': 12608,
 'padmé': 52362,
 "'bleedmedry": 52363,
 'pahalniuk': 52364,
 'honduras': 52365,
 'bursting': 10786,
 "pablo's": 41465,
 'irremediably': 52367,
 'presages': 40963,
 'bowlegged': 57832,
 'dalip': 65183,
 'entering': 6260,
 'newsradio': 76172,
 'presaged': 54150,
 "giallo's": 27663,
 'bouyant': 40964,
 'amerterish': 52368,
 'rajni': 18523,
 'leeves': 30610,
 'macauley': 34767,
 'seriously': 612,
 'sugercoma': 52369,
 'grimstead': 52370,
 "'fairy'": 52371,
 'zenda': 30611,
 "'twins'": 52372,
 'realisation': 17640,
 'highsmith': 27664,
 'raunchy': 7817,
 'incentives': 40965,
 'flatson': 52374,
 'snooker': 35097,
 'crazies': 16829,
 'crazier': 14902,
 'grandma': 7094,
 'napunsaktha': 52375,
 'workmanship': 30612,
 'reisner': 52376,
 "sanford's": 61306,
 '\x91doña': 52377,
 'modest': 6108,
 "everything's": 19153,
 'hamer': 40966,
 "couldn't'": 52379,
 'quibble': 13001,
 'socking': 52380,
 'tingler': 21931,
 'gutman': 52381,
 'lachlan': 40967,
 'tableaus': 52382,
 'headbanger': 52383,
 'spoken': 2847,
 'cerebrally': 34768,
 "'road": 23490,
 'tableaux': 21932,
 "proust's": 40968,
 'periodical': 40969,
 "shoveller's": 52385,
 'tamara': 25263,
 'affords': 17641,
 'concert': 3249,
 "yara's": 87955,
 'someome': 52386,
 'lingering': 8424,
 "abraham's": 41511,
 'beesley': 34769,
 'cherbourg': 34770,
 'kagan': 28624,
 'snatch': 9097,
 "miyazaki's": 9260,
 'absorbs': 25264,
 "koltai's": 40970,
 'tingled': 64027,
 'crossroads': 19511,
 'rehab': 16121,
 'falworth': 52389,
 'sequals': 52390,
 ...}

• 단어 인덱스 딕셔너리 -> 인덱스 단어 딕셔너리로 역변환

index_word = dict([(value, key) for (key, value) in word_index.items()])
index_word

{34701: 'fawn',
 52006: 'tsukino',
 52007: 'nunnery',
 16816: 'sonja',
 63951: 'vani',
 1408: 'woods',
 16115: 'spiders',
 2345: 'hanging',
 2289: 'woody',
 52008: 'trawling',
 52009: "hold's",
 11307: 'comically',
 40830: 'localized',
 30568: 'disobeying',
 52010: "'royale",
 40831: "harpo's",
 52011: 'canet',
 19313: 'aileen',
 52012: 'acurately',
 52013: "diplomat's",
 25242: 'rickman',
 6746: 'arranged',
 52014: 'rumbustious',
 52015: 'familiarness',
 52016: "spider'",
 68804: 'hahahah',
 52017: "wood'",
 40833: 'transvestism',
 34702: "hangin'",
 2338: 'bringing',
 40834: 'seamier',
 34703: 'wooded',
 52018: 'bravora',
 16817: 'grueling',
 1636: 'wooden',
 16818: 'wednesday',
 52019: "'prix",
 34704: 'altagracia',
 52020: 'circuitry',
 11585: 'crotch',
 57766: 'busybody',
 52021: "tart'n'tangy",
 14129: 'burgade',
 52023: 'thrace',
 11038: "tom's",
 52025: 'snuggles',
 29114: 'francesco',
 52027: 'complainers',
 52125: 'templarios',
 40835: '272',
 52028: '273',
 52130: 'zaniacs',
 34706: '275',
 27631: 'consenting',
 40836: 'snuggled',
 15492: 'inanimate',
 52030: 'uality',
 11926: 'bronte',
 4010: 'errors',
 3230: 'dialogs',
 52031: "yomada's",
 34707: "madman's",
 30585: 'dialoge',
 52033: 'usenet',
 40837: 'videodrome',
 26338: "kid'",
 52034: 'pawed',
 30569: "'girlfriend'",
 52035: "'pleasure",
 52036: "'reloaded'",
 40839: "kazakos'",
 52037: 'rocque',
 52038: 'mailings',
 11927: 'brainwashed',
 16819: 'mcanally',
 52039: "tom''",
 25243: 'kurupt',
 21905: 'affiliated',
 52040: 'babaganoosh',
 40840: "noe's",
 40841: 'quart',
 359: 'kids',
 5034: 'uplifting',
 7093: 'controversy',
 21906: 'kida',
 23379: 'kidd',
 52041: "error'",
 52042: 'neurologist',
 18510: 'spotty',
 30570: 'cobblers',
 9878: 'projection',
 40842: 'fastforwarding',
 52043: 'sters',
 52044: "eggar's",
 52045: 'etherything',
 40843: 'gateshead',
 34708: 'airball',
 25244: 'unsinkable',
 7180: 'stern',
 52046: "cervi's",
 40844: 'dnd',
 11586: 'dna',
 20598: 'insecurity',
 52047: "'reboot'",
 11037: 'trelkovsky',
 52048: 'jaekel',
 52049: 'sidebars',
 52050: "sforza's",
 17633: 'distortions',
 52051: 'mutinies',
 30602: 'sermons',
 40846: '7ft',
 52052: 'boobage',
 52053: "o'bannon's",
 23380: 'populations',
 52054: 'chulak',
 27633: 'mesmerize',
 52055: 'quinnell',
 10307: 'yahoo',
 52057: 'meteorologist',
 42577: 'beswick',
 15493: 'boorman',
 40847: 'voicework',
 52058: "ster'",
 22922: 'blustering',
 52059: 'hj',
 27634: 'intake',
 5621: 'morally',
 40849: 'jumbling',
 52060: 'bowersock',
 52061: "'porky's'",
 16821: 'gershon',
 40850: 'ludicrosity',
 52062: 'coprophilia',
 40851: 'expressively',
 19500: "india's",
 34710: "post's",
 52063: 'wana',
 5283: 'wang',
 30571: 'wand',
 25245: 'wane',
 52321: 'edgeways',
 34711: 'titanium',
 40852: 'pinta',
 178: 'want',
 30572: 'pinto',
 52065: 'whoopdedoodles',
 21908: 'tchaikovsky',
 2103: 'travel',
 52066: "'victory'",
 11928: 'copious',
 22433: 'gouge',
 52067: "chapters'",
 6702: 'barbra',
 30573: 'uselessness',
 52068: "wan'",
 27635: 'assimilated',
 16116: 'petiot',
 52069: 'most\x85and',
 3930: 'dinosaurs',
 352: 'wrong',
 52070: 'seda',
 52071: 'stollen',
 34712: 'sentencing',
 40853: 'ouroboros',
 40854: 'assimilates',
 40855: 'colorfully',
 27636: 'glenne',
 52072: 'dongen',
 4760: 'subplots',
 52073: 'kiloton',
 23381: 'chandon',
 34713: "effect'",
 27637: 'snugly',
 40856: 'kuei',
 9092: 'welcomed',
 30071: 'dishonor',
 52075: 'concurrence',
 23382: 'stoicism',
 14896: "guys'",
 52077: "beroemd'",
 6703: 'butcher',
 40857: "melfi's",
 30623: 'aargh',
 20599: 'playhouse',
 11308: 'wickedly',
 1180: 'fit',
 52078: 'labratory',
 40859: 'lifeline',
 1927: 'screaming',
 4287: 'fix',
 52079: 'cineliterate',
 52080: 'fic',
 52081: 'fia',
 34714: 'fig',
 52082: 'fmvs',
 52083: 'fie',
 52084: 'reentered',
 30574: 'fin',
 52085: 'doctresses',
 52086: 'fil',
 12606: 'zucker',
 31931: 'ached',
 52088: 'counsil',
 52089: 'paterfamilias',
 13885: 'songwriter',
 34715: 'shivam',
 9654: 'hurting',
 299: 'effects',
 52090: 'slauther',
 52091: "'flame'",
 52092: 'sommerset',
 52093: 'interwhined',
 27638: 'whacking',
 52094: 'bartok',
 8775: 'barton',
 21909: 'frewer',
 52095: "fi'",
 6192: 'ingrid',
 30575: 'stribor',
 52096: 'approporiately',
 52097: 'wobblyhand',
 52098: 'tantalisingly',
 52099: 'ankylosaurus',
 17634: 'parasites',
 52100: 'childen',
 52101: "jenkins'",
 52102: 'metafiction',
 17635: 'golem',
 40860: 'indiscretion',
 23383: "reeves'",
 57781: "inamorata's",
 52104: 'brittannica',
 7916: 'adapt',
 30576: "russo's",
 48246: 'guitarists',
 10553: 'abbott',
 40861: 'abbots',
 17649: 'lanisha',
 40863: 'magickal',
 52105: 'mattter',
 52106: "'willy",
 34716: 'pumpkins',
 52107: 'stuntpeople',
 30577: 'estimate',
 40864: 'ugghhh',
 11309: 'gameplay',
 52108: "wern't",
 40865: "n'sync",
 16117: 'sickeningly',
 40866: 'chiara',
 4011: 'disturbed',
 40867: 'portmanteau',
 52109: 'ineffectively',
 82143: "duchonvey's",
 37519: "nasty'",
 1285: 'purpose',
 52112: 'lazers',
 28105: 'lightened',
 52113: 'kaliganj',
 52114: 'popularism',
 18511: "damme's",
 30578: 'stylistics',
 52115: 'mindgaming',
 46449: 'spoilerish',
 52117: "'corny'",
 34718: 'boerner',
 6792: 'olds',
 52118: 'bakelite',
 27639: 'renovated',
 27640: 'forrester',
 52119: "lumiere's",
 52024: 'gaskets',
 884: 'needed',
 34719: 'smight',
 1297: 'master',
 25905: "edie's",
 40868: 'seeber',
 52120: 'hiya',
 52121: 'fuzziness',
 14897: 'genesis',
 12607: 'rewards',
 30579: 'enthrall',
 40869: "'about",
 52122: "recollection's",
 11039: 'mutilated',
 52123: 'fatherlands',
 52124: "fischer's",
 5399: 'positively',
 34705: '270',
 34720: 'ahmed',
 9836: 'zatoichi',
 13886: 'bannister',
 52127: 'anniversaries',
 30580: "helm's",
 52128: "'work'",
 34721: 'exclaimed',
 52129: "'unfunny'",
 52029: '274',
 544: 'feeling',
 52131: "wanda's",
 33266: 'dolan',
 52133: '278',
 52134: 'peacoat',
 40870: 'brawny',
 40871: 'mishra',
 40872: 'worlders',
 52135: 'protags',
 52136: 'skullcap',
 57596: 'dastagir',
 5622: 'affairs',
 7799: 'wholesome',
 52137: 'hymen',
 25246: 'paramedics',
 52138: 'unpersons',
 52139: 'heavyarms',
 52140: 'affaire',
 52141: 'coulisses',
 40873: 'hymer',
 52142: 'kremlin',
 30581: 'shipments',
 52143: 'pixilated',
 30582: "'00s",
 18512: 'diminishing',
 1357: 'cinematic',
 14898: 'resonates',
 40874: 'simplify',
 40875: "nature'",
 40876: 'temptresses',
 16822: 'reverence',
 19502: 'resonated',
 34722: 'dailey',
 52144: '2\x85',
 27641: 'treize',
 52145: 'majo',
 21910: 'kiya',
 52146: 'woolnough',
 39797: 'thanatos',
 35731: 'sandoval',
 40879: 'dorama',
 52147: "o'shaughnessy",
 4988: 'tech',
 32018: 'fugitives',
 30583: 'teck',
 76125: "'e'",
 40881: 'doesn’t',
 52149: 'purged',
 657: 'saying',
 41095: "martians'",
 23418: 'norliss',
 27642: 'dickey',
 52152: 'dicker',
 52153: "'sependipity",
 8422: 'padded',
 57792: 'ordell',
 40882: "sturges'",
 52154: 'independentcritics',
 5745: 'tempted',
 34724: "atkinson's",
 25247: 'hounded',
 52155: 'apace',
 15494: 'clicked',
 30584: "'humor'",
 17177: "martino's",
 52156: "'supporting",
 52032: 'warmongering',
 34725: "zemeckis's",
 21911: 'lube',
 52157: 'shocky',
 7476: 'plate',
 40883: 'plata',
 40884: 'sturgess',
 40885: "nerds'",
 20600: 'plato',
 34726: 'plath',
 40886: 'platt',
 52159: 'mcnab',
 27643: 'clumsiness',
 3899: 'altogether',
 42584: 'massacring',
 52160: 'bicenntinial',
 40887: 'skaal',
 14360: 'droning',
 8776: 'lds',
 21912: 'jaguar',
 34727: "cale's",
 1777: 'nicely',
 4588: 'mummy',
 18513: "lot's",
 10086: 'patch',
 50202: 'kerkhof',
 52161: "leader's",
 27644: "'movie",
 52162: 'uncomfirmed',
 40888: 'heirloom',
 47360: 'wrangle',
 52163: 'emotion\x85',
 52164: "'stargate'",
 40889: 'pinoy',
 40890: 'conchatta',
 41128: 'broeke',
 40891: 'advisedly',
 17636: "barker's",
 52166: 'descours',
 772: 'lots',
 9259: 'lotr',
 9879: 'irs',
 52167: 'lott',
 40892: 'xvi',
 34728: 'irk',
 52168: 'irl',
 6887: 'ira',
 21913: 'belzer',
 52169: 'irc',
 27645: 'ire',
 40893: 'requisites',
 7693: 'discipline',
 52961: 'lyoko',
 11310: 'extend',
 873: 'nature',
 52170: "'dickie'",
 40894: 'optimist',
 30586: 'lapping',
 3900: 'superficial',
 52171: 'vestment',
 2823: 'extent',
 52172: 'tendons',
 52173: "heller's",
 52174: 'quagmires',
 52175: 'miyako',
 20601: 'moocow',
 52176: "coles'",
 40895: 'lookit',
 52177: 'ravenously',
 40896: 'levitating',
 52178: 'perfunctorily',
 30587: 'lookin',
 40898: "lot'",
 52179: 'lookie',
 34870: 'fearlessly',
 52181: 'libyan',
 40899: 'fondles',
 35714: 'gopher',
 40901: 'wearying',
 52182: "nz's",
 27646: 'minuses',
 52183: 'puposelessly',
 52184: 'shandling',
 31268: 'decapitates',
 11929: 'humming',
 40902: "'nother",
 21914: 'smackdown',
 30588: 'underdone',
 40903: 'frf',
 52185: 'triviality',
 25248: 'fro',
 8777: 'bothers',
 52186: "'kensington",
 73: 'much',
 34730: 'muco',
 22615: 'wiseguy',
 27648: "richie's",
 40904: 'tonino',
 52187: 'unleavened',
 11587: 'fry',
 40905: "'tv'",
 40906: 'toning',
 14361: 'obese',
 30589: 'sensationalized',
 40907: 'spiv',
 6259: 'spit',
 7364: 'arkin',
 21915: 'charleton',
 16823: 'jeon',
 21916: 'boardroom',
 4989: 'doubts',
 3084: 'spin',
 53083: 'hepo',
 27649: 'wildcat',
 10584: 'venoms',
 52191: 'misconstrues',
 18514: 'mesmerising',
 40908: 'misconstrued',
 52192: 'rescinds',
 52193: 'prostrate',
 40909: 'majid',
 16479: 'climbed',
 34731: 'canoeing',
 52195: 'majin',
 57804: 'animie',
 40910: 'sylke',
 14899: 'conditioned',
 40911: 'waddell',
 52196: '3\x85',
 41188: 'hyperdrive',
 34732: 'conditioner',
 53153: 'bricklayer',
 2576: 'hong',
 52198: 'memoriam',
 30592: 'inventively',
 25249: "levant's",
 20638: 'portobello',
 52200: 'remand',
 19504: 'mummified',
 27650: 'honk',
 19505: 'spews',
 40912: 'visitations',
 52201: 'mummifies',
 25250: 'cavanaugh',
 23385: 'zeon',
 40913: "jungle's",
 34733: 'viertel',
 27651: 'frenchmen',
 52202: 'torpedoes',
 52203: 'schlessinger',
 34734: 'torpedoed',
 69876: 'blister',
 52204: 'cinefest',
 34735: 'furlough',
 52205: 'mainsequence',
 40914: 'mentors',
 9094: 'academic',
 20602: 'stillness',
 40915: 'academia',
 52206: 'lonelier',
 52207: 'nibby',
 52208: "losers'",
 40916: 'cineastes',
 4449: 'corporate',
 40917: 'massaging',
 30593: 'bellow',
 19506: 'absurdities',
 53241: 'expetations',
 40918: 'nyfiken',
 75638: 'mehras',
 52209: 'lasse',
 52210: 'visability',
 33946: 'militarily',
 52211: "elder'",
 19023: 'gainsbourg',
 20603: 'hah',
 13420: 'hai',
 34736: 'haj',
 25251: 'hak',
 4311: 'hal',
 4892: 'ham',
 53259: 'duffer',
 52213: 'haa',
 66: 'had',
 11930: 'advancement',
 16825: 'hag',
 25252: "hand'",
 13421: 'hay',
 20604: 'mcnamara',
 52214: "mozart's",
 30731: 'duffel',
 30594: 'haq',
 13887: 'har',
 44: 'has',
 2401: 'hat',
 40919: 'hav',
 30595: 'haw',
 52215: 'figtings',
 15495: 'elders',
 52216: 'underpanted',
 52217: 'pninson',
 27652: 'unequivocally',
 23673: "barbara's",
 52219: "bello'",
 12997: 'indicative',
 40920: 'yawnfest',
 52220: 'hexploitation',
 52221: "loder's",
 27653: 'sleuthing',
 32622: "justin's",
 52222: "'ball",
 52223: "'summer",
 34935: "'demons'",
 52225: "mormon's",
 34737: "laughton's",
 52226: 'debell',
 39724: 'shipyard',
 30597: 'unabashedly',
 40401: 'disks',
 2290: 'crowd',
 10087: 'crowe',
 56434: "vancouver's",
 34738: 'mosques',
 6627: 'crown',
 52227: 'culpas',
 27654: 'crows',
 53344: 'surrell',
 52229: 'flowless',
 52230: 'sheirk',
 40923: "'three",
 52231: "peterson'",
 52232: 'ooverall',
 40924: 'perchance',
 1321: 'bottom',
 53363: 'chabert',
 52233: 'sneha',
 13888: 'inhuman',
 52234: 'ichii',
 52235: 'ursla',
 30598: 'completly',
 40925: 'moviedom',
 52236: 'raddick',
 51995: 'brundage',
 40926: 'brigades',
 1181: 'starring',
 52237: "'goal'",
 52238: 'caskets',
 52239: 'willcock',
 52240: "threesome's",
 52241: "mosque'",
 52242: "cover's",
 17637: 'spaceships',
 40927: 'anomalous',
 27655: 'ptsd',
 52243: 'shirdan',
 21962: 'obscenity',
 30599: 'lemmings',
 30600: 'duccio',
 52244: "levene's",
 52245: "'gorby'",
 25255: "teenager's",
 5340: 'marshall',
 9095: 'honeymoon',
 3231: 'shoots',
 12258: 'despised',
 52246: 'okabasho',
 8289: 'fabric',
 18515: 'cannavale',
 3537: 'raped',
 52247: "tutt's",
 17638: 'grasping',
 18516: 'despises',
 40928: "thief's",
 8926: 'rapes',
 52248: 'raper',
 27656: "eyre'",
 52249: 'walchek',
 23386: "elmo's",
 40929: 'perfumes',
 21918: 'spurting',
 52250: "exposition'\x85",
 52251: 'denoting',
 34740: 'thesaurus',
 40930: "shoot'",
 49759: 'bonejack',
 52253: 'simpsonian',
 30601: 'hebetude',
 34741: "hallow's",
 52254: 'desperation\x85',
 34742: 'incinerator',
 10308: 'congratulations',
 52255: 'humbled',
 5924: "else's",
 40845: 'trelkovski',
 52256: "rape'",
 59386: "'chapters'",
 52257: '1600s',
 7253: 'martian',
 25256: 'nicest',
 52259: 'eyred',
 9457: 'passenger',
 6041: 'disgrace',
 52260: 'moderne',
 5120: 'barrymore',
 52261: 'yankovich',
 40931: 'moderns',
 52262: 'studliest',
 52263: 'bedsheet',
 14900: 'decapitation',
 52264: 'slurring',
 52265: "'nunsploitation'",
 34743: "'character'",
 9880: 'cambodia',
 52266: 'rebelious',
 27657: 'pasadena',
 40932: 'crowne',
 52267: "'bedchamber",
 52268: 'conjectural',
 52269: 'appologize',
 52270: 'halfassing',
 57816: 'paycheque',
 20606: 'palms',
 52271: "'islands",
 40933: 'hawked',
 21919: 'palme',
 40934: 'conservatively',
 64007: 'larp',
 5558: 'palma',
 21920: 'smelling',
 12998: 'aragorn',
 52272: 'hawker',
 52273: 'hawkes',
 3975: 'explosions',
 8059: 'loren',
 52274: "pyle's",
 6704: 'shootout',
 18517: "mike's",
 52275: "driscoll's",
 40935: 'cogsworth',
 52276: "britian's",
 34744: 'childs',
 52277: "portrait's",
 3626: 'chain',
 2497: 'whoever',
 52278: 'puttered',
 52279: 'childe',
 52280: 'maywether',
 3036: 'chair',
 52281: "rance's",
 34745: 'machu',
 4517: 'ballet',
 34746: 'grapples',
 76152: 'summerize',
 30603: 'freelance',
 52283: "andrea's",
 52284: '\x91very',
 45879: 'coolidge',
 18518: 'mache',
 52285: 'balled',
 40937: 'grappled',
 18519: 'macha',
 21921: 'underlining',
 5623: 'macho',
 19507: 'oversight',
 25257: 'machi',
 11311: 'verbally',
 21922: 'tenacious',
 40938: 'windshields',
 18557: 'paychecks',
 3396: 'jerk',
 11931: "good'",
 34748: 'prancer',
 21923: 'prances',
 52286: 'olympus',
 21924: 'lark',
 10785: 'embark',
 7365: 'gloomy',
 52287: 'jehaan',
 52288: 'turaqui',
 20607: "child'",
 2894: 'locked',
 52289: 'pranced',
 2588: 'exact',
 52290: 'unattuned',
 783: 'minute',
 16118: 'skewed',
 40940: 'hodgins',
 34749: 'skewer',
 52291: 'think\x85',
 38765: 'rosenstein',
 52292: 'helmit',
 34750: 'wrestlemanias',
 16826: 'hindered',
 30604: "martha's",
 52293: 'cheree',
 52294: "pluckin'",
 40941: 'ogles',
 11932: 'heavyweight',
 82190: 'aada',
 11312: 'chopping',
 61534: 'strongboy',
 41342: 'hegemonic',
 40942: 'adorns',
 41346: 'xxth',
 34751: 'nobuhiro',
 52298: 'capitães',
 52299: 'kavogianni',
 13422: 'antwerp',
 6538: 'celebrated',
 52300: 'roarke',
 40943: 'baggins',
 31270: 'cheeseburgers',
 52301: 'matras',
 52302: "nineties'",
 52303: "'craig'",
 12999: 'celebrates',
 3383: 'unintentionally',
 14362: 'drafted',
 52304: 'climby',
 52305: '303',
 18520: 'oldies',
 9096: 'climbs',
 9655: 'honour',
 34752: 'plucking',
 30074: '305',
 5514: 'address',
 40944: 'menjou',
 42592: "'freak'",
 19508: 'dwindling',
 9458: 'benson',
 52307: 'white’s',
 40945: 'shamelessness',
 21925: 'impacted',
 52308: 'upatz',
 3840: 'cusack',
 37567: "flavia's",
 52309: 'effette',
 34753: 'influx',
 52310: 'boooooooo',
 52311: 'dimitrova',
 13423: 'houseman',
 25259: 'bigas',
 52312: 'boylen',
 52313: 'phillipenes',
 40946: 'fakery',
 27658: "grandpa's",
 27659: 'darnell',
 19509: 'undergone',
 52315: 'handbags',
 21926: 'perished',
 37778: 'pooped',
 27660: 'vigour',
 3627: 'opposed',
 52316: 'etude',
 11799: "caine's",
 52317: 'doozers',
 34754: 'photojournals',
 52318: 'perishes',
 34755: 'constrains',
 40948: 'migenes',
 30605: 'consoled',
 16827: 'alastair',
 52319: 'wvs',
 52320: 'ooooooh',
 34756: 'approving',
 40949: 'consoles',
 52064: 'disparagement',
 52322: 'futureistic',
 52323: 'rebounding',
 52324: "'date",
 52325: 'gregoire',
 21927: 'rutherford',
 34757: 'americanised',
 82196: 'novikov',
 1042: 'following',
 34758: 'munroe',
 52326: "morita'",
 52327: 'christenssen',
 23106: 'oatmeal',
 25260: 'fossey',
 40950: 'livered',
 13000: 'listens',
 76164: "'marci",
 52330: "otis's",
 23387: 'thanking',
 16019: 'maude',
 34759: 'extensions',
 52332: 'ameteurish',
 52333: "commender's",
 27661: 'agricultural',
 4518: 'convincingly',
 17639: 'fueled',
 54014: 'mahattan',
 40952: "paris's",
 52336: 'vulkan',
 52337: 'stapes',
 52338: 'odysessy',
 12259: 'harmon',
 4252: 'surfing',
 23494: 'halloran',
 49580: 'unbelieveably',
 52339: "'offed'",
 30607: 'quadrant',
 19510: 'inhabiting',
 34760: 'nebbish',
 40953: 'forebears',
 34761: 'skirmish',
 52340: 'ocassionally',
 52341: "'resist",
 21928: 'impactful',
 52342: 'spicier',
 40954: 'touristy',
 52343: "'football'",
 40955: 'webpage',
 52345: 'exurbia',
 52346: 'jucier',
 14901: 'professors',
 34762: 'structuring',
 30608: 'jig',
 40956: 'overlord',
 25261: 'disconnect',
 82201: 'sniffle',
 40957: 'slimeball',
 40958: 'jia',
 16828: 'milked',
 40959: 'banjoes',
 1237: 'jim',
 52348: 'workforces',
 52349: 'jip',
 52350: 'rotweiller',
 34763: 'mundaneness',
 52351: "'ninja'",
 11040: "dead'",
 40960: "cipriani's",
 20608: 'modestly',
 52352: "professor'",
 40961: 'shacked',
 34764: 'bashful',
 23388: 'sorter',
 16120: 'overpowering',
 18521: 'workmanlike',
 27662: 'henpecked',
 18522: 'sorted',
 52354: "jōb's",
 52355: "'always",
 34765: "'baptists",
 52356: 'dreamcatchers',
 52357: "'silence'",
 21929: 'hickory',
 52358: 'fun\x97yet',
 52359: 'breakumentary',
 15496: 'didn',
 52360: 'didi',
 52361: 'pealing',
 40962: 'dispite',
 25262: "italy's",
 21930: 'instability',
 6539: 'quarter',
 12608: 'quartet',
 52362: 'padmé',
 52363: "'bleedmedry",
 52364: 'pahalniuk',
 52365: 'honduras',
 10786: 'bursting',
 41465: "pablo's",
 52367: 'irremediably',
 40963: 'presages',
 57832: 'bowlegged',
 65183: 'dalip',
 6260: 'entering',
 76172: 'newsradio',
 54150: 'presaged',
 27663: "giallo's",
 40964: 'bouyant',
 52368: 'amerterish',
 18523: 'rajni',
 30610: 'leeves',
 34767: 'macauley',
 612: 'seriously',
 52369: 'sugercoma',
 52370: 'grimstead',
 52371: "'fairy'",
 30611: 'zenda',
 52372: "'twins'",
 17640: 'realisation',
 27664: 'highsmith',
 7817: 'raunchy',
 40965: 'incentives',
 52374: 'flatson',
 35097: 'snooker',
 16829: 'crazies',
 14902: 'crazier',
 7094: 'grandma',
 52375: 'napunsaktha',
 30612: 'workmanship',
 52376: 'reisner',
 61306: "sanford's",
 52377: '\x91doña',
 6108: 'modest',
 19153: "everything's",
 40966: 'hamer',
 52379: "couldn't'",
 13001: 'quibble',
 52380: 'socking',
 21931: 'tingler',
 52381: 'gutman',
 40967: 'lachlan',
 52382: 'tableaus',
 52383: 'headbanger',
 2847: 'spoken',
 34768: 'cerebrally',
 23490: "'road",
 21932: 'tableaux',
 40968: "proust's",
 40969: 'periodical',
 52385: "shoveller's",
 25263: 'tamara',
 17641: 'affords',
 3249: 'concert',
 87955: "yara's",
 52386: 'someome',
 8424: 'lingering',
 41511: "abraham's",
 34769: 'beesley',
 34770: 'cherbourg',
 28624: 'kagan',
 9097: 'snatch',
 9260: "miyazaki's",
 25264: 'absorbs',
 40970: "koltai's",
 64027: 'tingled',
 19511: 'crossroads',
 16121: 'rehab',
 52389: 'falworth',
 52390: 'sequals',
 ...}

• imdb.get_word_index()를 통해 얻은 데이터 인덱스
  • 단어 출현 빈도 기준 내림차순 정렬
  • index_word 값을 바꿔보면 -> 출현 빈도순 단어 살펴보기 가능

# 1순위
index_word[1]

# 2순위
index_word[2]

...

Q. 25번째 단어를 키로 지정, word_index에 어떤 Value가 있는지 확인

word_25th = index_word[25]
print(word_25th)

• train_data[0] 각 인덱스 매핑 단어들로 연결 -> 1개의 리뷰로 만들기

review = ' '.join([str(i) for i in train_data[0]])
review

'1 14 22 16 43 530 973 1622 1385 65 458 4468 66 3941 4 173 36 256 5 25 100 43 838 112 50 670 2 9 35 480 284 5 150 4 172 112 167 2 336 385 39 4 172 4536 1111 17 546 38 13 447 4 192 50 16 6 147 2025 19 14 22 4 1920 4613 469 4 22 71 87 12 16 43 530 38 76 15 13 1247 4 22 17 515 17 12 16 626 18 2 5 62 386 12 8 316 8 106 5 4 2223 5244 16 480 66 3785 33 4 130 12 16 38 619 5 25 124 51 36 135 48 25 1415 33 6 22 12 215 28 77 52 5 14 407 16 82 2 8 4 107 117 5952 15 256 4 2 7 3766 5 723 36 71 43 530 476 26 400 317 46 7 4 2 1029 13 104 88 4 381 15 297 98 32 2071 56 26 141 6 194 7486 18 4 226 22 21 134 476 26 480 5 144 30 5535 18 51 36 28 224 92 25 104 4 226 65 16 38 1334 88 12 16 283 5 16 4472 113 103 32 15 16 5345 19 178 32'

review = ' '.join([index_word.get(i-3, '?') for i in train_data[0]])
review

"? this film was just brilliant casting location scenery story direction everyone's really suited the part they played and you could just imagine being there robert ? is an amazing actor and now the same being director ? father came from the same scottish island as myself so i loved the fact there was a real connection with this film the witty remarks throughout the film were great it was just brilliant so much that i bought the film as soon as it was released for ? and would recommend it to everyone to watch and the fly fishing was amazing really cried at the end it was so sad and you know what they say if you cry at a film it must have been good and this definitely was also ? to the two little boy's that played the ? of norman and paul they were just brilliant children are often left out of the ? list i think because the stars that play them all grown up are such a big profile for the whole film but these children are amazing and should be praised for what they have done don't you think the whole story was so lovely because it was true and was someone's life after all that was shared with us all"

• 원-핫 인코딩
  • 10000 차원을 갖는 원-핫 인코딩으로 IMDB 데이터 처리

def one_hot_encoding(data, dim=10000): # imdb 데이터의 num_words를 10000으로 설정해서 dim도 10000으로 맞춰줍니다.
  results = np.zeros((len(data), dim))
  for i, d in enumerate(data):
    results[i, d] = 1.
  return results

x_train = one_hot_encoding(train_data)
x_test = one_hot_encoding(test_data)

print(x_train[0])

• 정수형 -> 실수형으로 변환

print(train_labels[0])
print(test_labels[0])

y_train = np.asarray(train_labels).astype('float32')
y_test = np.asarray(test_labels).astype('float32')

print(y_train[0])
print(y_test[0])

모델 구성

• Sequential()
  • 첫번째 레이어
    • Dense : 유닛수 16개
    • input_shape : (10000, )
    • activation : relu
    • name(레이어 이름) : input
  • 두번째 레이어
    • Dense : 유닛수 16개
    • activation : relu
    • name(레이어 이름) : hidden
  • 세번째 레이어
    • Dense 사용 -> 긍정/부정 결과만 받을 수 있게 유닛수는 1개
    • activation : sigmoid
    • name(레이어 이름) : output

import tensorflow as tf
from tensorflow.keras import models, layers

model = models.Sequential()
model.add(layers.Dense(16, activation='relu', input_shape=(10000, ), name='input'))
model.add(layers.Dense(16, activation='relu', name='hidden'))
model.add(layers.Dense(1, activation='sigmoid', name='output'))

모델 컴파일 & 학습

• optimizer : rmsprop
• loss : binary_crossentropy(∵긍정/부정 분류)
• metrics : accuracy

model.compile(optimizer='rmsprop',
              loss='binary_crossentropy',
              metrics=['accuracy'])
model.summary()

• 딥러닝 모델 학습
  • 학습 데이터 : x_train, y_train
  • epochs : 20
  • batch_size : 512
  • 테스트 데이터 : x_test, y_test

history = model.fit(x_train, y_train,
                    epochs=20,
                    batch_size=512,
                    validation_data=(x_test, y_test))

Epoch 1/20
49/49 [==============================] - 8s 122ms/step - loss: 0.4458 - accuracy: 0.8264 - val_loss: 0.3471 - val_accuracy: 0.8704
Epoch 2/20
49/49 [==============================] - 1s 18ms/step - loss: 0.2519 - accuracy: 0.9116 - val_loss: 0.2916 - val_accuracy: 0.8846
Epoch 3/20
49/49 [==============================] - 1s 18ms/step - loss: 0.1970 - accuracy: 0.9283 - val_loss: 0.2831 - val_accuracy: 0.8876
Epoch 4/20
49/49 [==============================] - 1s 18ms/step - loss: 0.1643 - accuracy: 0.9426 - val_loss: 0.3033 - val_accuracy: 0.8816
Epoch 5/20
49/49 [==============================] - 1s 17ms/step - loss: 0.1457 - accuracy: 0.9491 - val_loss: 0.3111 - val_accuracy: 0.8794
Epoch 6/20
49/49 [==============================] - 1s 18ms/step - loss: 0.1257 - accuracy: 0.9572 - val_loss: 0.3398 - val_accuracy: 0.8741
Epoch 7/20
49/49 [==============================] - 1s 18ms/step - loss: 0.1108 - accuracy: 0.9628 - val_loss: 0.3692 - val_accuracy: 0.8705
Epoch 8/20
49/49 [==============================] - 1s 18ms/step - loss: 0.0986 - accuracy: 0.9674 - val_loss: 0.3972 - val_accuracy: 0.8651
Epoch 9/20
49/49 [==============================] - 1s 17ms/step - loss: 0.0895 - accuracy: 0.9710 - val_loss: 0.4187 - val_accuracy: 0.8634
Epoch 10/20
49/49 [==============================] - 1s 18ms/step - loss: 0.0784 - accuracy: 0.9735 - val_loss: 0.4810 - val_accuracy: 0.8520
Epoch 11/20
49/49 [==============================] - 1s 19ms/step - loss: 0.0693 - accuracy: 0.9782 - val_loss: 0.4636 - val_accuracy: 0.8620
Epoch 12/20
49/49 [==============================] - 1s 21ms/step - loss: 0.0612 - accuracy: 0.9821 - val_loss: 0.5248 - val_accuracy: 0.8516
Epoch 13/20
49/49 [==============================] - 1s 19ms/step - loss: 0.0531 - accuracy: 0.9850 - val_loss: 0.5239 - val_accuracy: 0.8580
Epoch 14/20
49/49 [==============================] - 1s 17ms/step - loss: 0.0482 - accuracy: 0.9860 - val_loss: 0.5636 - val_accuracy: 0.8576
Epoch 15/20
49/49 [==============================] - 1s 18ms/step - loss: 0.0430 - accuracy: 0.9872 - val_loss: 0.5901 - val_accuracy: 0.8561
Epoch 16/20
49/49 [==============================] - 1s 19ms/step - loss: 0.0347 - accuracy: 0.9909 - val_loss: 0.6601 - val_accuracy: 0.8500
Epoch 17/20
49/49 [==============================] - 1s 18ms/step - loss: 0.0321 - accuracy: 0.9910 - val_loss: 0.6656 - val_accuracy: 0.8506
Epoch 18/20
49/49 [==============================] - 1s 17ms/step - loss: 0.0263 - accuracy: 0.9931 - val_loss: 0.7486 - val_accuracy: 0.8489
Epoch 19/20
49/49 [==============================] - 1s 18ms/step - loss: 0.0233 - accuracy: 0.9936 - val_loss: 0.7412 - val_accuracy: 0.8511
Epoch 20/20
49/49 [==============================] - 1s 17ms/step - loss: 0.0196 - accuracy: 0.9952 - val_loss: 0.7769 - val_accuracy: 0.8479

• 결과 시각화

import matplotlib.pyplot as plt

history_dict = history.history

loss = history_dict['loss']
val_loss = history_dict['val_loss']
epochs = range(1, len(loss) + 1)

fig = plt.figure(figsize=(12, 5))

ax1 = fig.add_subplot(1, 2, 1)
ax1.plot(epochs, loss, color='blue', label='train_loss')
ax1.plot(epochs, val_loss, color='red', label='val_loss')
ax1.set_title('Train and Validation Loss')
ax1.set_xlabel('Epochs')
ax1.set_ylabel('Loss')
ax1.grid()
ax1.legend()

accuracy = history_dict['accuracy']
val_accuracy = history_dict['val_accuracy']

ax2 = fig.add_subplot(1, 2, 2)
ax2.plot(epochs, accuracy, color='blue', label='train_accuracy')
ax2.plot(epochs, val_accuracy, color='red', label='val_accuracy')
ax2.set_title('Train and Validation Accuracy')
ax2.set_xlabel('Epochs')
ax2.set_ylabel('Accuracy')
ax2.grid()
ax2.legend()

plt.show()

분석

• val_loss 증가 추세
• val_accuracy 감소 추세
• OverFitting(과대적합)

8-5. 마무리하며

---

Q. IMDB 데이터셋을 이용한 딥러닝 분류 모델은 과대적합 -> 과대적합 방지를 위해 적용해볼 것은?

A. 다양한 학습 데이터 수집&학습, 파라미터가 적은 모델의 선택, 학습 데이터 특성 수 줄이기 등 ➡️ 모델 단순화 작업을 시도해볼 것 같음

• 에폭 수 줄이기
• 학습률 or 모델 크기 줄여보기
• 비슷한 데이터 추가 수집 등..