ML Kaggle (basic level)

Question 1

тянуть,

вытащить

Answer 1

pull,

pull out

Question 2

польза

Answer 2

utility [ютилити]

Question 3

мера качества модели

Answer 3

measure of model quality

Question 4

точность модели

Answer 4

model accuracy

Question 5

удобство

Answer 5

convenience

Question 6

узел,

листовые узлы

Answer 6

node,

leaf nodes [лиф ноудс]

Question 7

глубокое дерево,
мелкое дерево

Answer 7

deep tree,
shallow tree

Question 8

Это проваливает (не в состоянии) охватить все различия в сырых данных

Answer 8

It fails to capture as many distinctions in the raw data

Question 9

complicated (усложненный)

Answer 9

sophisticated [сэфистикейтэд]

Question 10

Это может привести к лучшей производительности

Answer 10

It can lead to better performance

Question 11

Рандомный лес обеспечивает гораздо большую предсказательную точность, чем одиночное древо решений

Answer 11

The Random Forest provides much better predictive accuracy than a single decision tree

Question 12

ключевые выводы (сущ из глагола take)

Answer 12

key takeaways

Question 13

настраивать

Answer 13

to tune

Question 14

статистическая техника (как метод)

Answer 14

statistical technique [тэкнИк]

Question 15

Признаки, ось X

Метка, ось y

Answer 15

features,

label

Question 16

снижаться, расти. Например функция растет, падает

Answer 16

to decrease, to increase

Question 17

Уравнение линейной регрессии и что это

Answer 17

Linear regression equation helps to find the relationship between x [экс] and y [уай] by fitting a straight [стрэйт] line that best represents the data

Question 18

Коэффициент наклона

Answer 18

slope coefficient [коэфишэнт]

Question 19

умноженный на

Answer 19

multiplied by

Question 20

ускорение, ускорить

Answer 20

acceleration [аксэлэрэйшэн],
to accelerate

Question 21

быть более устойчивым к выбросам

Answer 21

be more robust to outliers

Question 22

синоним to tune (настраивать, регулировать)

Answer 22

to adjust

Question 23

прилагательное: постепенно, с повторениями с повторением итераций

Answer 23

iteratively

Question 24

Градиентный спуск и что это

Answer 24

Gradient descent is a mathematical technique that iteratively finds the best weights and bias that produce the model with the lowest loss

Question 25

Модель сошлась (типо стала рабочей и тд)

Answer 25

The model has converged

Question 26

Сходимость (модели на пример)

Answer 26

model convergence [конвёрджэнс]

Question 27

Функции потерь для линейных моделей всегда создают выпуклую поверхность

Answer 27

The loss functions for linear models always produce a convex surface

Question 28

свойство

Answer 28

property

Question 29

Как результат, в результате

Answer 29

as a result

Question 30

выпуклость функции потерь

Answer 30

convexity of loss function

Question 31

Кривая потерь и что она показывает

Answer 31

The loss curve [кёрв] illustrate how loss changes over iterations during training

Question 32

Что такое скорость обучения модели и какова цель при выборе скорости

Answer 32

Learning rate is a floating point number, that you set, and that influences how quickly the model converges. The goal is to pick a learning rate that’s not too high and not too low, so that the model converges quickly

Question 33

Размер выборки для обновления весов. И объяснить что это.

Answer 33

Batch Size is a hyperparameter that defines how many examples the model processes before updating its weights and bias.

Question 34

Что такое гиперпараметры и параметры. Key Differences

Answer 34

Hyperparameters are values, you choose before training (learning rate, batch size, epochs, etc.). Parameters are values the model learns during training (weights, bias)

Question 35

Две распространненные техники определения бача, для градиентного спуска. И их объяснение.

Answer 35

There are two common gradient descent techniques: 1. Stochastic [стокхастик] Gradient Descent (SGD) - uses only one example per iteration (batch size = 1) - Works with enough iterations but introduces noise (fluctuations in loss) - Stochastic means that the single example is chosen randomly 2. Mini-Batch Stochastic Gradient Descent (Mini-Batch SGD) - A compromise between full-batch and SGD - Uses a small subset(подмножество) of example per batch (1

Question 36

Определение маленького и большого размера батча

Answer 36

1. Small batch size (1-32) behaves like SGD, more noise, but learns faster. 2. Large batch size (128+) behaves like full-batch gradient descent, more stable but requires more memory.

Question 37

Что такое эпоха?

Answer 37

The Epoch [ипок] is a hyperparameter, that is one full pass though the entire training dataset. One epoch is not enough for the model to learn meaningful patterns (значимые закономерности)

Question 38

Что такое логистическая регрессия?

Answer 38

Logistic regression is a powerful technique for estimating probabilities. It‘s commonly used in classification problems where the output needs to be a probability between 0 and 1.

Question 39

How does logistic regression work?

Answer 39

It takes an input (features) and applies (применять) a sigmoid function to transform it into a probability.

Question 40

What is Sigmoid Function?

Answer 40

The sigmoid function ensures (provides) that the model’s output is always between 0 and 1, making it suitable [suitable] (fitting, corresponding) for probability estimation.

Question 41

Обучение модели логистической регрессии, два основных отличение от линейной

Answer 41

Training Logistic Regression Models is the same process as training linear regression models, with two key distinctions: 1. Log [лог] Loss as the loss function. 2. Regularisation [рэгуларайзэйшн] to prevent overfitting.

Question 42

Почему мы используем log loss для logistic regression?

Answer 42

Squared loss works well for a linear model where the rate of change (темп изменения) is constant. However, the rate of change of a logistic regression model is not constant. The sigmoid curve is s-shaped rather than linear.

Question 43

Точность положительных предсказаний

Answer 43

precision

Question 44

Что такое классификация?

Answer 44

Classification is the task of predicting witch of a set of classes (к какому из множества классов) an example belongs to

Question 45

Порог классификации

Answer 45

The classification threshold [трэшолд] is used to convert a probability estimate into a definitive class. (окончательный класс)

Question 46

1) если предсказанное значение выше порога 2) если предсказанное значение ниже порога

Answer 46

1) If the predicted probability is above the threshold. Then the example is assigned to the positive class. 2) If the predicted probability is below the threshold. Then the example is assigned to the negative class.

Question 47

Более высокий и более низкий порог

Answer 47

The choice of classification threshold significantly (существенно) impacts the model’s performance. A higher threshold increases precision (повышает точность) but may lower recall (меньше полнота класса) A lower threshold increases recall but may lower precision

Question 48

Для чего используется «матрица путаницы»?

Answer 48

A confusion matrix [мэйтрикс] is used to evaluate the performance of a binary classifier

Question 49

Как матрица путаницы делит предсказания

Answer 49

A confusion matrix compares the model’s predictions with the actual ground truth, breaking them down into four possible outcomes: True positive, true negative and false positive, false negative

Question 50

Метрики гипотетической идеальной модели классификации

Answer 50

A hypothetical [хайпотэтикал] perfect classification model should have 0 False Positive and 0 False Negative predictions. If a model achieves this, the performance metrics would be: Accuracy (точность всех предсказаний) is equal 1 Recall (полнота правильных примеров) is equal 1 False Positive Rate (количество ложных примеров) is equal 0 Precision (точность положительных предсказаний) is equal 1

Question 51

Что такое Receiver-operating characteristic (ROC) curve?

Answer 51

ROC [ароси] curve shows how a model’s TPR [типиар] and FPR [эфпиар] change across thresholds (в зависимости от пороговых значений)

Question 52

Что обозначает area under the ROC curve?

Answer 52

AUC [эй б си] measures the overall performance of a classifier.

Question 53

От чего зависит оптимальный порог классификации?

Answer 53

The optimal threshold depends on the trade-off (компромисс) between false positives and false negatives