14  Prediction

Prerequisites

Key concepts and skills

Software and packages

library(arrow)
library(nflverse)
library(poissonreg)
library(tidymodels)
library(tidyverse)
library(tinytable)

14.1 Introduction

As discussed in Chapter 12, models tend to be focused on either inference or prediction. There are, in general, different cultures depending on your focus. One reason for this is a different emphasis of causality, which will be introduced in Chapter 15. I am talking very generally here, but often with inference we will be very concerned about causality, and with prediction we will be less so. That means the quality of our predictions will break-down when conditions are quite different from what our model was expecting—but how do we know when conditions are sufficiently different for us to be worried?

Another way for this cultural difference is because the rise of data science and machine learning in particular, has been substantially driven by the development of models in Python by people with a computer science or engineering background. This means there is an additional vocabulary difference because much of inference came out of statistics. Again, this is all speaking very broadly.

In this chapter, I begin with a focus on prediction using the R approach of tidymodels. I then introduce one of those grey areas—and the reason that I have been trying to speak broadly—lasso regression. That was developed by statisticians, but is mostly used for prediction. Finally, I introduce all of this in Python.

14.2 Prediction with tidymodels

14.2.1 Linear models

When we are focused on prediction, we will often want to fit many models. One way to do this is to copy and paste code many times. This is okay, and it is the way that most people get started but it is prone to making errors that are hard to find. A better approach will:

  1. scale more easily;
  2. enable us to think carefully about over-fitting; and
  3. add model evaluation.

The use of tidymodels (Kuhn and Wickham 2020) satisfies these criteria by providing a coherent grammar that allows us to easily fit a variety of models. Like the tidyverse, it is a package of packages.

By way of illustration, we want to estimate the following model for the simulated running data:

\[ \begin{aligned} y_i | \mu_i &\sim \mbox{Normal}(\mu_i, \sigma) \\ \mu_i &= \beta_0 +\beta_1x_i \end{aligned} \]

where \(y_i\) refers to the marathon time of some individual \(i\) and \(x_i\) refers to their five-kilometer time. Here we say that the marathon time of some individual \(i\) is normally distributed with a mean of \(\mu\) and a standard deviation of \(\sigma\), where the mean depends on two parameters \(\beta_0\) and \(\beta_1\) and their five-kilometer time. Here “~” means “is distributed as”. We use this slightly different notation from earlier to be more explicit about the distributions being used, but this model is equivalent to \(y_i=\beta_0+\beta_1 x_i + \epsilon_i\), where \(\epsilon\) is normally distributed.

As we are focused on prediction, we are worried about over-fitting our data, which would limit our ability to make claims about other datasets. One way to partially address this is to split our dataset in two using initial_split().

sim_run_data <- 
  read_parquet(file = here::here("outputs/data/running_data.parquet"))

set.seed(853)

sim_run_data_split <-
  initial_split(
    data = sim_run_data,
    prop = 0.80
  )

sim_run_data_split
<Training/Testing/Total>
<160/40/200>

Having split the data, we then create training and test datasets with training() and testing().

sim_run_data_train <- training(sim_run_data_split)

sim_run_data_test <- testing(sim_run_data_split)

We have placed 80 per cent of our dataset into the training dataset. We will use that to estimate the parameters of our model. We have kept the remaining 20 per cent of it back, and we will use that to evaluate our model. Why might we do this? Our concern is the bias-variance trade-off, which haunts all aspects of modeling. We are concerned that our results may be too particular to the dataset that we have, such that they are not applicable to other datasets. To take an extreme example, consider a dataset with ten observations. We could come up with a model that perfectly hits those observations. But when we took that model to other datasets, even those generated by the same underlying process, it would not be accurate.

One way to deal with this concern is to split the data in this way. We use the training data to inform our estimates of the coefficients, and then use the test data to evaluate the model. A model that too closely matched the data in the training data would not do well in the test data, because it would be too specific to the training data. The use of this test-training split enables us the opportunity to build an appropriate model.

It is more difficult to do this separation appropriately than one might initially think. We want to avoid the situation where aspects of the test dataset are present in the training dataset because this inappropriately telegraphs what is about to happen. This is called data leakage. But if we consider data cleaning and preparation, which likely involves the entire dataset, it may be that some features of each are influencing each other. Kapoor and Narayanan (2023) find extensive data leakage in applications of machine learning that could invalidate much research.

To use tidymodels we first specify that we are interested in linear regression with linear_reg(). We then specify the type of linear regression, in this case multiple linear regression, with set_engine(). Finally, we specify the model with fit(). While this requires considerably more infrastructure than the base R approach detailed above, the advantage of this approach is that it can be used to fit many models; we have created a model factory, as it were.

sim_run_data_first_model_tidymodels <-
  linear_reg() |>
  set_engine(engine = "lm") |>
  fit(
    marathon_time ~ five_km_time + was_raining,
    data = sim_run_data_train
  )

The estimated coefficients are summarized in the first column of Table 12.4. For instance, we find that on average in our dataset, five-kilometer run times that are one minute longer are associated with marathon times that are about eight minutes longer.

14.2.2 Logistic regression

We can also use tidymodels for logistic regression problems. To accomplish this, we first need to change the class of our dependent variable into a factor because this is required for classification models.

week_or_weekday <- 
  read_parquet(file = "outputs/data/week_or_weekday.parquet")

set.seed(853)

week_or_weekday <-
  week_or_weekday |>
  mutate(is_weekday = as_factor(is_weekday))

week_or_weekday_split <- initial_split(week_or_weekday, prop = 0.80)
week_or_weekday_train <- training(week_or_weekday_split)
week_or_weekday_test <- testing(week_or_weekday_split)

week_or_weekday_tidymodels <-
  logistic_reg(mode = "classification") |>
  set_engine("glm") |>
  fit(
    is_weekday ~ number_of_cars,
    data = week_or_weekday_train
  )

As before, we can make a graph of the actual results compared with our estimates. But one nice aspect of this is that we could use our test dataset to evaluate our model’s prediction ability more thoroughly, for instance through a confusion matrix, which specifies the count of each prediction by what the truth was. We find that the model does well on the held-out dataset. There were 90 observations where the model predicted it was a weekday, and it was actually a weekday, and 95 where the model predicted it was a weekend, and it was a weekend. It was wrong for 15 observations, and these were split across seven where it predicted a weekday, but it was a weekend, and eight where it was the opposite case.

week_or_weekday_tidymodels |>
  predict(new_data = week_or_weekday_test) |>
  cbind(week_or_weekday_test) |>
  conf_mat(truth = is_weekday, estimate = .pred_class)
          Truth
Prediction  0  1
         0 90  8
         1  7 95

14.2.2.1 US political support

One approach is to use tidymodels to build a prediction-focused logistic regression model in the same way as before, i.e. a validation set approach (James et al. [2013] 2021, 176). In this case, the probability will be that of voting for Biden.

ces2020 <- 
  read_parquet(file = "outputs/data/ces2020.parquet")

set.seed(853)

ces2020_split <- initial_split(ces2020, prop = 0.80)
ces2020_train <- training(ces2020_split)
ces2020_test <- testing(ces2020_split)

ces_tidymodels <-
  logistic_reg(mode = "classification") |>
  set_engine("glm") |>
  fit(
    voted_for ~ gender + education,
    data = ces2020_train
  )

ces_tidymodels
parsnip model object


Call:  stats::glm(formula = voted_for ~ gender + education, family = stats::binomial, 
    data = data)

Coefficients:
                  (Intercept)                     genderMale  
                       0.2157                        -0.4697  
educationHigh school graduate          educationSome college  
                      -0.1857                         0.3502  
              education2-year                education4-year  
                       0.2311                         0.6700  
           educationPost-grad  
                       0.9898  

Degrees of Freedom: 34842 Total (i.e. Null);  34836 Residual
Null Deviance:      47000 
Residual Deviance: 45430    AIC: 45440

And then evaluate it on the test set. It appears as though the model is having difficulty identifying Trump supporters.

ces_tidymodels |>
  predict(new_data = ces2020_test) |>
  cbind(ces2020_test) |>
  conf_mat(truth = voted_for, estimate = .pred_class)
          Truth
Prediction Trump Biden
     Trump   656   519
     Biden  2834  4702

When we introduced tidymodels, we discussed the importance of randomly constructing training and test sets. We use the training dataset to estimate parameters, and then evaluate the model on the test set. It is natural to ask why we should be subject to the whims of randomness and whether we are making the most of our data. For instance, what if a good model is poorly evaluated because of some random inclusion in the test set? Further, what if we do not have a large test set?

One commonly used resampling method that goes some way to addressing this is \(k\)-fold cross-validation. In this approach we create \(k\) different samples, or “folds”, from the dataset without replacement. We then fit the model to the first \(k-1\) folds, and evaluate it on the last fold. We do this \(k\) times, once for every fold, such that every observation will be used for training \(k-1\) times and for testing once. The \(k\)-fold cross-validation estimate is then the average mean squared error (James et al. [2013] 2021, 181). For instance, vfold_cv() from tidymodels can be used to create, say, ten folds.

set.seed(853)

ces2020_10_folds <- vfold_cv(ces2020, v = 10)

The model can then be fit across the different combinations of folds with fit_resamples(). In this case, the model will be fit ten times.

ces2020_cross_validation <-
  fit_resamples(
    object = logistic_reg(mode = "classification") |> set_engine("glm"),
    preprocessor = recipe(voted_for ~ gender + education,
                          data = ces2020),
    resamples = ces2020_10_folds,
    metrics = metric_set(accuracy, sens),
    control = control_resamples(save_pred = TRUE)
  )

We might be interested to understand the performance of our model and we can use collect_metrics() to aggregate them across the folds (?tbl-metricsvoters-1). These types of details would typically be mentioned in passing in the main content of the paper, but included in great detail in an appendix. The average accuracy of our model across the folds is 0.61, while the average sensitivity is 0.19 and the average specificity is 0.90.

collect_metrics(ces2020_cross_validation) |>
  select(.metric, mean) |>
  tt() |> 
  style_tt(j = 1:2, align = "lr") |> 
  format_tt(digits = 2, num_mark_big = ",", num_fmt = "decimal") |> 
  setNames(c("Metric", "Mean"))
conf_mat_resampled(ces2020_cross_validation) |>
  mutate(Proportion = Freq / sum(Freq)) |>
  tt() |> 
  style_tt(j = 1:4, align = "llrr") |> 
  format_tt(digits = 2, num_mark_big = ",", num_fmt = "decimal")
Table 14.1: Average metrics across the ten folds of a logistic regression to predict voter preference
Metric Mean
accuracy 0.61
sens 0.19
Prediction Truth Freq Proportion
Trump Trump 327.5 0.08
Trump Biden 267.7 0.06
Biden Trump 1,428.3 0.33
Biden Biden 2,331.9 0.54

What does this mean? Accuracy is the proportion of observations that were correctly classified. The result of 0.61 suggests the model is doing better than a coin toss, but not much more. Sensitivity is the proportion of true observations that are identified as true (James et al. [2013] 2021, 145). In this case that would mean the model predicted a respondent voted for Trump and they did. Specificity is the proportion of false observations that are identified as false (James et al. [2013] 2021, 145). In this case it is the proportion of voters that voted for Biden, that were predicted to vote for Biden. This confirms our initial thought that the model is having trouble identifying Trump supporters.

We can see this in more detail by looking at the confusion matrix (?tbl-metricsvoters-2). When used with resampling approaches, such as cross-validation, the confusion matrix is computed for each fold and then averaged. The model is predicting Biden much more than we might expect from our knowledge of how close the 2020 election was. It suggests that our model may need additional variables to do a better job.

Finally, it may be the case that we are interested in individual-level results, and we can add these to our dataset with collect_predictions().

ces2020_with_predictions <-
  cbind(
    ces2020,
    collect_predictions(ces2020_cross_validation) |>
      arrange(.row) |>
      select(.pred_class)
  ) |>
  as_tibble()

For instance, we can see that the model is essentially predicting support for Biden for all individuals apart from males with no high school, high school graduates, or two years of college (Table 14.2).

ces2020_with_predictions |>
  group_by(gender, education, voted_for) |>
  count(.pred_class) |>
  tt() |> 
  style_tt(j = 1:5, align = "llllr") |> 
  format_tt(digits = 0, num_mark_big = ",", num_fmt = "decimal") |> 
  setNames(c(
      "Gender",
      "Education",
      "Voted for",
      "Predicted vote",
      "Number"
    ))
Table 14.2: The model is predicting support for Biden for all females, and for many males, regardless of education
Gender Education Voted for Predicted vote Number
Female No HS Trump Biden 206
Female No HS Biden Biden 228
Female High school graduate Trump Biden 3,204
Female High school graduate Biden Biden 3,028
Female Some college Trump Biden 1,842
Female Some college Biden Biden 3,325
Female 2-year Trump Biden 1,117
Female 2-year Biden Biden 1,739
Female 4-year Trump Biden 1,721
Female 4-year Biden Biden 4,295
Female Post-grad Trump Biden 745
Female Post-grad Biden Biden 2,853
Male No HS Trump Trump 132
Male No HS Biden Trump 123
Male High school graduate Trump Trump 2,054
Male High school graduate Biden Trump 1,528
Male Some college Trump Biden 1,992
Male Some college Biden Biden 2,131
Male 2-year Trump Trump 1,089
Male 2-year Biden Trump 1,026
Male 4-year Trump Biden 2,208
Male 4-year Biden Biden 3,294
Male Post-grad Trump Biden 1,248
Male Post-grad Biden Biden 2,426

14.2.3 Poisson regression

We can use tidymodels to estimate Poisson models with poissonreg (Kuhn and Frick 2022) (Table 13.4).

count_of_A <- 
  read_parquet(file = "outputs/data/count_of_A.parquet")

set.seed(853)

count_of_A_split <-
  initial_split(count_of_A, prop = 0.80)
count_of_A_train <- training(count_of_A_split)
count_of_A_test <- testing(count_of_A_split)

grades_tidymodels <-
  poisson_reg(mode = "regression") |>
  set_engine("glm") |>
  fit(
    number_of_As ~ department,
    data = count_of_A_train
  )

The results of this estimation are in the second column of Table 13.4. They are similar to the estimates from glm(), but the number of observations is less because of the split.

14.3 Lasso regression

Shoulders of giants

Dr Robert Tibshirani is Professor in the Departments of Statistics and Biomedical Data Science at Stanford University. After earning a PhD in Statistics from Stanford University in 1981, he joined the University of Toronto as an assistant professor. He was promoted to full professor in 1994 and moved to Stanford in 1998. He made fundamental contributions including GAMs, mentioned above, and lasso regression, which is a way of automated variable selection. He is an author of James et al. ([2013] 2021). He was awarded the COPSS Presidents’ Award in 1996 and was appointed a Fellow of the Royal Society in 2019.

14.4 Prediction with Python

14.4.1 Setup

We will use Python within VSCode, which is a free IDE from Microsoft that you can download here. You then install the Quarto and Python extensions.

14.4.2 Data

Read in data using parquet.

Manipulate using pandas

14.4.3 Model

14.4.3.1 scikit-learn

14.4.3.2 TensorFlow

14.5 Exercises

Practice

  1. (Plan) Consider the following scenario: Each day for a year your uncle and you play darts. Each round consists of throwing three darts each. At the end of each round you add the points that your three darts hit. So if you hit 3, 5, and 10, then your total score for that round is 18. Your uncle is somewhat benevolent, and if you hit a number less than 5, he pretends not to see that, allowing you the chance to re-throw that dart. Pretend that each day you play 15 rounds. Please sketch out what that dataset could look like and then sketch a graph that you could build to show all observations.
  2. (Simulate) Please further consider the scenario described and simulate the situation. Compare your uncle’s total with your total if you didn’t get the chance to re-throw, and the total that actually end up with. Please include at least ten tests based on the simulated data.
  3. (Acquire) Please describe one possible source of such a dataset (or an equivalent sport or situation of interest to you).
  4. (Explore) Please use ggplot2 to build the graph that you sketched. Then use tidymodels to build a forecasting model of who wins.
  5. (Communicate) Please write two paragraphs about what you did.

Quiz

Class activities

Task

Please use nflverse to load some statistics for NFL quarterbacks during the regular season. The data dictionary will be useful to make sense of the data.

qb_regular_season_stats <- 
  load_player_stats(seasons = TRUE) |> 
  filter(season_type == "REG" & position == "QB")

Pretend that you are an NFL analyst and that it is half way through the 2023 regular season, i.e. Week 9 has just finished. I am interested in the best forecasting model of passing_epa that you can generate for each team in the remainder of the season, i.e. Weeks 10-18.

Use Quarto, and include an appropriate title, author, date, link to a GitHub repo, sections, and citations, and write a 2-3 page report for management. The best performance will likely require creative feature engineering. You are welcome to use R or Python, any model, but you should be careful to specify the model and explain, at a high-level, how it works. Be careful about leakage!