아래의 이미지처럼 Clone or download -> Download ZIP 버튼을 누르셔서 자료를 받으실 수 있습니다.
혹은 다음과 같이 git clone 커맨드로 받으실 수 있습니다.
git clone https://github.com/realblack0/ai_lecture_fundamental.git
찾기(ctrl + F) 기능으로 키워드를 검색하면 해당 내용이 있는 필기자료를 찾을 수 있습니다.
- 사용 라이브러리
- 주제
- 함수, 클래스, 속성 등 라이브러리가 제공하는 기능 명칭
- numpy
- Fancy Indexing, Broadcasting, Universal Function, Factory Method, tensorflow의 MNIST 데이터셋 열어보기
- shape, ndim, size, dtype, itemsize, arange, zeros, ones, full, _like, empty, linsapce, matrix, flat, copy, view, linalg, split, vsplit, hsplit, reshape, resize, stack, vstack, hstack, concatenate, dstack, r_, c_, s_, ix_, newaxis
- numpy
- fashion mnist 데이터셋 열어보기, IMDB 데이터 열어보기
- nditer, finished, iternext, flatten, ravel
- numpy, pickle, pandas
- Serialize
- loadtxt, genfromtxt, read_csv
- pandas
- structured array, fancy indexing, tips 데이터 분석
- info, head, tail, sample, describe, skew, kurt
- pandas, pandas-profiling
- 값 뽑기, 인덱스 조작, 멀티인덱스, Data Type, EDA
- rename, loc, iloc, at, iat, set_index, reset_index, astype, corr, drop, duplecated, drop_duplicates
- pandas
- boolean indexing, predicate, str 메소드
- isin, where, mask, query, startswith, endswith, in, split, strip, replace, concatenate, join, groupby, stack, unstack
- pandas, fbprophet
- aggregation, map 시리즈, Time Series(시계열 분석)
- groupby, pivot_table, crosstab, map, apply, applymap, calender, time, datetime, to_datetime, resample, date_range, Period
- pandas
- Tidy Data, 추천시스템
- concat, join, merge, append, melt, wide_to_long
- pandas, folium
- 추천시스템, 가짜 코로나19 지도
- matplotlib, pdvega, folium
- State Machine 방식, OOP 방식, style 변경, interactive 그래프, 한글 폰트 설정, GeoJson, TopoJson, Choropleth(단계구분도)
- scikit-learn, folium
- KNN, MLP, Choropleth(단계구분도)
- scikit-learn
- 기계학습, 지도학습, random, hold out
- train_test_split, randint, shuffle, seed
- scikit-learn, mglearn
- 정형데이터, 비정형데이터, regression을 classification으로 치환, imbalanced data, 차원의 저주, data leakage, cross validation
- cut, qcut, score, cross_val_score, KFold, dummy
- scikit-learn, seaborn, sklearn-evaluation
- hold out, learning curve
- boxplot, violinplot, boxenplot, StratifiedKFold, Group KFold, ShuffleSplit, cross_val_predict, corss_validate, learning_curve
- scikit-learn, missingno, pandas
- preprocessing(전처리), scaling(정규화), encoding, missing data(결측값)
- StandardScaler, MinMaxScaler, LabelEncoder, get_dummies, OneHotEncoder, fit_transform, inverse_transform, isnull, isna, fillna
- scikit-learn, seaborn
- Kaggle, binary vs multi class, one vs one, one vs rest, multi label, linear vs non-linear, decision boundary, probability vs non-probability, deterministic vs stochastic, chaining
- FacetGrid, make_blobs, OneVsOneClassifier, OneVsRestClassifier, MultiOutputClassifier, GaussianNB, predict_proba, predict_log_proba
- scikit-learn
- 모델 선택 기준, 오캄의 면도칼, No Free Lunch, Pipeline, 하이퍼 파라미터 찾기, 3way hold-out, metric, TP, FP, TN, FN, precision, recall, f1-score
- make_classification, Pipeline, make_pipeline, GridSearchCV, classification_report, confusion_matrix
- scikit-learn
- SVM, support vector machine, 알고리즘과 하이퍼파라미터 동시에 찾기, feature selection
- make_circle, SVC, SVR, validation_curve, RandomizedSearchCV, ParameterGrid, FunctionTransformer, @np.vectorize, Binarizer, SelectKBest, chi2
- scikit-learn
- 다중공선성, impute, feature selection(filter, wrapper, embedded), CART, parametric vs non-parametric, ensemble, GridSearchCV의 내부 구조, VotingClassifier + Pipeline + GridSearchCV
- KNNImputer, SimpleImputer, chi2, SelectKBest, f_regression, RFE, insert, DecisionTreeRegressor, VotingClassifier
- scikit-learn, surprise, xgboost
- 추천시스템, Ensemble(앙상블), bagging(배깅), boosting(부스팅), stacking, unsupervised learning(비지도 학습)
- BaggingClassifier, RandomForestClassifier, AdaBoostClassifier, StackingClassifier, KMeans, DBSCAN
- scikit-learn, scipy
- unsupervised learning(비지도 학습), Agglomerative Clustering(병합 군집화), Hierachical Clustering(계층적 군집화), Dendrogram, Principle Component Analysis(주성분분석), LDA(잠재 디리클레 할당), t-sne, 딥러닝의 역사
- DBSCAN, KMeans, AgglomerativeClustering, linkage, dendrogram, PCA, LatentDirichletAllocation
- tensorflow, keras
- 딥러닝 프레임워크, activation function, loss function, optimizer, metrics
- Sequential, Dense, add, summary, compile, fit, epochs
- tensorflow, keras, scikit-learn
- mnist, fashion mnist, scaling, activation function, advance activation, history, learning curve, validation curve
- Flatten, predict, relu, softmax
- tensorflow, keras, tensorflow_hub, tensorflow_datasets
- backpropagation, feedforward, batch size, Tensor, keras : functional paradigm, transfer learning
- constant, Model, Input
- tensorflow, keras, tensorflow_datasets
- optimizer, zero padding, feature extraction, transfer learning, convolution
- get_word_index, pad_sequence, take, trainable
- tensorflow , keras
- mse, mae, callback, 혼자서 공부하는 법
- get_file, Callback, TensorBoard, EarlyStopping, History, layers, get_layer, weights, bias, units
- tensorflow, keras
- learning curve graph interprete, overfitting, underfitting, weight decay, drop out, tf.data, wrappers
- kernel_regularizer, Dropout, KerasClassifier, KerasRegressor, Dataset, feature_column, DenseFeatures
- tensorflow, keras, tpot
- estimator, 모델 저장 및 로드, model zoo, 문자열 formatting, HDF5, tpot, AutoML
- from_tensor_slise, take, shuffle, repeat, batch, DNNClassifier, ModelCheckpoint, save_weights_only, save, save_weight, load_model, load_weights, format, TPOTCLassifier, TPOTRegressor, export
- tensorflow, keras
- CNN (Convolution Neural Network)
- Convolution2D, Conv2D, MaxPolling2D, MaxPool2D, plot_model
- tensorflow
- feature map visualization, filter visualization, Augmentation, compile에 문자열 대신 객체 쓰기
- ImageDataGenerator, flow, flow_from_dataframe, flow_from_directory, Dropout
- tensorflow, scikit-learn
- keras와 scikit-learn 연동하기
- KerasClassifier, learning_curve, GridSearchCV
- issues page: https://github.com/realblack0/ai_lecture_fundamental/issues
- e-mail: realblack0@gmail.com
- AI 이노베이션 스퀘어: https://ai.koipa.or.kr/
- LMS 홈페이지: https://lms.koipa.or.kr/