Data-Driven Transfer Learning Framework for Estimating Turning Movement Counts

Hyunsoo Noh; James Tokishi; Ryan Hatch; Xiaobo Ma; Zepu Wang

REVIEW 4 major objections 5 minor 3 references

Data-Driven Transfer Learning Framework for Estimating Turning Movement Counts

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A transfer learning model that borrows data from similar intersections estimates turning movement counts with lower errors than eight standard regression models.

desk verdict A plausible first application of transfer learning to turning movement counts, undermined by a confounded evaluation: the TL model gets target-feature access during source selection while the baselines don't. read the letter →

arxiv 2412.09861 v1 pith:GNHI2JQ3 submitted 2024-12-13 cs.LG

Xiaobo Ma , Hyunsoo Noh , Ryan Hatch , James Tokishi , Zepu Wang This is my paper · ORCID

classification cs.LG

keywords transferlearningturningmovementcountstrafficflowestimationTrAdaBoost.R2controllereventdatapoint-of-interestlassoregressionintersection

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Turning movement counts (left, through, right turns) at signalized intersections are usually collected by sensors or manual counting, which is costly to scale. This paper proposes a transfer learning framework that estimates those counts for intersections without sensors by reusing data from instrumented 'source' intersections. The framework selects important traffic variables with Lasso regression, matches each target intersection to a similar source intersection by Pearson correlation, picks the most similar 10% of source instances by cosine similarity, and fits a Two-stage TrAdaBoost.R2 model. On 30 Tucson intersections, the model reports the lowest average MAE and RMSE for all three movement types compared with eight baseline regression models. If the result holds, traffic agencies could obtain turning movement counts at unmonitored intersections from existing controller event data, road geometry, and points-of-interest data.

What carries the argument

The load-bearing mechanism is Two-stage TrAdaBoost.R2 (TrA), a boosting procedure that begins with a combined set of source and selected source instances, assigns equal initial weights, and at each round reweights source instances by their adjusted error while keeping target-like instances' weights controlled by a pre-specified schedule, finally outputting the weak estimator with the lowest cross-validation error. Before boosting, the framework identifies comparable intersections by maximizing the sum of Pearson correlations between the same Lasso-selected traffic variables of source and target intersections (Eq. 10), then substitutes the top 10% of source instances ranked by cosine similarity to the target instances as pseudo-target labeled data. This chain carries the argument: feature selection, similarity matching, instance substitution, and TrA reweighting together let labeled source data stand in for missing target labels.

What would settle it

Re-run the exact TrA pipeline on the same 30 intersections, but replace the Pearson-correlation intersection matching and the cosine-similarity top-10% selection with random matching and random selection of the same sizes; if the average MAE remains close to 9.57/28.59/10.25 and RMSE close to 13.65/40.62/13.98, the similarity mechanism is not doing the work and the central claim fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that an instance-based transfer learning model, Two-stage TrAdaBoost.R2, can estimate turning movement counts at a target intersection using no observed target labels, by reweighting source-domain instances that resemble the target and by first matching the target to a similar source intersection. It reports average test errors on 30 Tucson intersections of MAE 9.57 for left turns, 28.59 for through movements, and 10.25 for right turns, with RMSE 13.65, 40.62, and 13.98, all lower than the corresponding errors of KNN, SVR, RF, MLP, AdaBoost, XGBoost, CatBoost, and LightGBM. The paper argues that the gains come from relaxing the assumption that source and target data distributions are identical, which lets the model build scene-specific estimators for intersections with different geometry, signal timing, and local behavior.

Load-bearing premise

The framework assumes that an intersection or a single data row that resembles the target on detector activity, lane counts, POI, and time will also have similar turning movement counts, so the labels of the most similar source rows can stand in for the missing target labels.

Editorial extensions

If this is right

If the central claim is right, turning movement counts can be estimated at intersections that lack loop detectors or cameras, using data already collected by traffic signal controllers.
A city could build a library of instrumented intersections and, for any new intersection, automatically select similar sources and obtain movement-specific estimates without field counting.
The reported accuracy gains imply that scene-specific models, rather than one global model, are the better default for TMC estimation across heterogeneous intersections.
Because the method uses controller event data plus road and POI features, it can be extended to new intersections quickly at low marginal cost.
The framework's reweighting mechanism could be re-run whenever new controller data arrive, giving a scalable path to continuous traffic monitoring.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper does not test whether its Pearson-correlation intersection matching and cosine-similarity instance selection beat random selection; a reader should treat the transferability premise as unproven until such an ablation is run.
Because the evaluation protocol trains on the other 29 intersections for each target, the source library always contains near-neighbors of the target; deployed systems with a smaller or less similar source pool would likely see smaller gains than the 9.57/28.59/10.25 MAE values reported.
The same instance-reweighting chain could be applied to other intersection-level quantities such as queue lengths, delays, or pedestrian volumes, but the similarity premise would need to be re-validated for each target quantity.
A direct comparison against a model fine-tuned with a small number of real target labels would clarify how much of the gain comes from transfer versus from the boosting procedure itself.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a legitimate first application of transfer learning to turning movement count estimation, using real controller event data from 30 Tucson intersections. The data work is real, and the comparison against eight baselines is extensive. But the main claim — that TL beats generic models — is not actually established by the numbers as reported, because the TL pipeline sees the target intersection's features when it selects which source data to use, and the baselines don't.

What's genuinely new here is applying the source-selection-and-substitution scheme to TMC data; I don't see that in the cited literature. The feature set (detector occupancy, green time, cycle counts, lane geometry, POI) is reasonable, and Lasso gives a sensible ranking. The evaluation is on real intersections, not simulations, and the reported error numbers are internally consistent.

The stress-test note is fair. The TL model computes Pearson correlation between each target and each source intersection (Eq. 10) and then uses cosine similarity to keep the top 10% of source instances that are closest to the target features (§3.4). That's transductive use of target inputs. Baselines are trained once on all 29 source intersections and applied to the target, with no analogous feature-based selection. So the comparison conflates transfer learning with target-aware data selection. An ablation with random source selection, or giving baselines the same matched subset, would be needed. As it stands, the reported 20-30% error reduction could come mostly from selecting rows that are structurally similar to the target, not from the TrAdaBoost.R2 mechanism itself.

Also missing: error bars or significance tests across the 30 holdout intersections, no description of hyperparameter tuning for the weak learner (the weak learner G is never specified), and no code/data release. Minor issues: the text spends a lot of space on generic introductions of Lasso and AdaBoost, and the claim to 'guarantee' high accuracy in the intro is overblown. None of this is fatal; it's fixable with a proper experimental design.

Who this is for: transportation researchers and practitioners working on TMC estimation or sensor-free traffic data; also a good teaching case for why transductive vs inductive baselines matter. It deserves a serious referee — the application is relevant, the data are real, and the gap is correctable. But the revision needs to fix the comparison, add significance tests, and ideally release code/data.

Referee Report

4 major / 5 minor

Summary. The paper proposes a data-driven transfer learning (TL) framework for estimating turning movement counts (TMCs) at signalized intersections. The pipeline combines Lasso regression for feature selection, a 'similar-intersection matching' step based on Pearson correlation between target and source intersection features (Eq. 10), a 'target domain data substitution' step that selects the top 10% of source instances by cosine similarity to the target features (Section 3.4), and a Two-stage TrAdaBoost.R2 (TrA) regression model. The framework is evaluated on 30 intersections in Tucson, Arizona, using a leave-one-out protocol, with event-based detector data, infrastructure data, and POI data. The reported results (Tables 3 and 4) show the TL model achieving the lowest average MAE and RMSE among nine models (eight baselines plus TL) for left-turn, through, and right-turn movements.

Significance. If the reported accuracy gains are real, the framework would be a practically valuable contribution: it estimates TMCs without additional physical sensors, using data already available to many traffic agencies, and it explicitly addresses domain shift across intersections. The paper merits credit for assembling a realistic multi-source dataset, implementing a complete TL pipeline, and performing a 30-intersection evaluation. However, the central claim of superiority is not yet established because the comparison is confounded: the TL method uses target-intersection input features for source selection and instance reweighting, while the baselines do not, and no statistical significance testing or ablation is provided. The transferability premise—that input-feature similarity implies label similarity—is load-bearing but unvalidated. The work is therefore promising but requires substantial additional analysis before the claims can be accepted.

major comments (4)

[Section 4.2.2 and Section 4.4] The comparison between the TL model and the eight baselines is confounded: the TL pipeline uses the target intersection's input features to select a matched source intersection (Eq. 10, Section 3.3) and to select the top 10% of source instances by cosine similarity (Section 3.4), then reweights those instances via TrA (Section 3.5). The baselines are trained on all 29 source intersections with no access to target features. Consequently, the reported MAE/RMSE differences do not isolate 'transfer learning'; they compare transductive target-feature-matched instance weighting against inductive training. To support the claim that TL is the cause of improvement, the authors should include ablations such as random source-instance selection, cosine-matched source instances without TrA reweighting, and a variant that gives the baselines the same matched subset. They should also report paired statistical tests (e.g., Wilcoxon signed-rank or paired t-test over the 30 holdout intersections) with effect sizes and error bars.
[Section 4.2.2, Section 4.4] The hyperparameter tuning procedure is not described relative to the leave-one-out (LOO) evaluation. The paper states that 'a grid search method was initially developed to identify the optimal hyperparameter values for baseline models and the proposed TL method' (Section 4.2.2), but it does not specify whether tuning is performed inside each of the 30 LOO folds or once on all data. If the target intersection's data participates in selecting the threshold θ (top 10%), the TrA parameters S and F, or the baseline hyperparameters, the reported test errors are optimistically biased. The authors must either describe a nested cross-validation scheme or state explicitly that all hyperparameters were fixed before LOO evaluation.
[Sections 3.3 and 3.4, Eq. (10)] The framework's central transferability assumption is that source intersections with high Pearson correlation on the selected input features, and source instances with high cosine similarity to target features, have similar turning movement counts. This assumption is load-bearing because the selected source instances' labels substitute for the unavailable target labels in Eq. (6). The paper does not validate this premise against random selection or alternative similarity measures, nor does it report any diagnostic showing that matched intersections/instances actually have more similar TMC values than unmatched ones. Without such a check, the reported accuracy gains could stem from the concentration of high-volume similar instances rather than from a meaningful correspondence of turning patterns.
[Section 4.4, Tables 3 and 4] Only point estimates of MAE and RMSE are reported—the averages across the 30 LOO test intersections. There are no confidence intervals, standard deviations, or significance tests. The paper's qualitative statements in Section 4.4, such as 'the TL model consistently shows the lowest MAE values' and 'exhibits the lowest and most consistent RMSE values', go beyond what the tables support. The authors should provide per-fold statistics and a formal comparison (e.g., paired tests across the 30 folds) so the reader can judge whether the differences are reliable given the small number of target intersections.

minor comments (5)

[Abstract and Section 1] The abstract contains a typo: 'intersections, Accurate TMCs' should be 'intersections. Accurate TMCs'. The phrase 'state-of-the-art regression models' appears in the abstract but the paper does not define what 'state-of-the-art' means in this context; listing the eight models in Section 4.2.2 is helpful, but the abstract should align with the actual evaluation.
[Section 4.2.1] The sentence 'On average, data from one to four days per intersection were utilized for model training and validation' is vague. It should specify how many 15-minute intervals per intersection are in the training and test sets, and how many total instances (n) are used. This information is necessary for interpreting the MAE/RMSE magnitudes.
[Section 4.2.2] The baseline 'LightBoost' is elsewhere called 'LightGBM' (Section 4.4 and Table captions). Please use a consistent name throughout.
[References] Reference [Dai Wenyuan, Y.Q., Guirong, X. and Yong, Y. (2007)] is a malformed duplicate of the Dai et al. (2007) entry. Please consolidate and correct the citation format.
[Section 3.5, Algorithm 1] The pseudocode for TrA is underspecified: the exact formula for the adjusted error e_i^t, the definition of β_t, and the stopping criterion in the AdaBoost.R2 call are not given. Since the proposed method relies on this algorithm, please provide the complete algorithm or a precise reference to the version used.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the claimed accuracy is an empirical result on held-out intersections, and the self-citations are not load-bearing.

full rationale

No circularity was found in the proposed derivation. The transfer learning pipeline (Sections 3.3-3.5) is a transductive instance-weighting procedure: it uses target-intersection input features to select and weight source instances via Pearson correlation (Eq. 10) and a cosine-similarity threshold (Section 3.4), then trains a standard TrAdaBoost.R2 model on source labels only. The reported MAE and RMSE values are computed on held-out target intersections (Section 4.4) without using target labels during training, so the central claim is not defined in terms of the quantity it predicts. The self-citations, such as X. Ma et al. (2024), are background references for transfer learning and do not supply the empirical result. The comparison with baselines may be confounded because baselines do not receive target-feature access for instance selection, and the Lasso variable selection and hyperparameter grid search are potential leakage concerns, but these are validity issues rather than circularity: no fitted parameter is renamed as a prediction, and no equation reduces to its own input by construction.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The framework rests on domain assumptions that controller event, infrastructure, and POI features can predict TMCs, and that feature similarity identifies transferable intersections and source instances. These assumptions are heuristic and not independently validated. Several free hyperparameters (Lasso lambda, similarity threshold, boosting steps, folds, weak learner) are fitted or chosen without reported values, but no entities are invented.

free parameters (5)

Lasso regularization parameter lambda
Controls sparsity of feature selection in Section 3.2; tuned via grid search per Section 4.2.2 but value not reported.
Cosine similarity threshold theta (top 10%) = top 10%
Selects D'_S used as pseudo-target data in Section 3.4; chosen as a hyperparameter, not derived.
TrA step count S = 10
Set in Section 3.5; arbitrary choice.
TrA cross-validation folds F = 5
Set in Section 3.5; arbitrary choice.
Weak learner G and its hyperparameters
The estimator G in TrA (Section 3.5) is never named; hyperparameters for baselines and TL are grid-searched but not reported.

assumptions (4)

domain assumption Turning movement counts can be estimated from traffic controller event data, intersection layout, and POI features.
Underlies the entire feature set in Table 1 and Eqs. (1)-(3); not separately validated.
ad hoc to paper Pearson correlation between the same features identifies source intersections with similar turning patterns to the target.
Similar-intersection matching in Section 3.3, Eq. (10), assumes this without validation.
ad hoc to paper The top 10% of source instances by cosine similarity to target features provide labels that can stand in for target labels.
Target domain data substitution in Section 3.4; a heuristic proxy for missing target labels.
standard math AdaBoost.R2 and the TrAdaBoost weight update rule improve regression transfer.
The boosting algorithm in Section 3.5 is taken from cited prior work (Drucker 1997; Dai et al. 2007).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Data-Driven Transfer Learning Framework for Estimating Turning Movement Counts." pith.science (2026). https://pith.science/paper/GNHI2JQ3

@misc{pith2026241209861,
  author       = {Pith},
  title        = {Pith review of: Data-Driven Transfer Learning Framework for Estimating Turning Movement Counts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GNHI2JQ3}},
  note         = {Machine review of arXiv:2412.09861}
}

read the original abstract

Urban transportation networks are vital for the efficient movement of people and goods, necessitating effective traffic management and planning. An integral part of traffic management is understanding the turning movement counts (TMCs) at intersections, Accurate TMCs at intersections are crucial for traffic signal control, congestion mitigation, and road safety. In general, TMCs are obtained using physical sensors installed at intersections, but this approach can be cost-prohibitive and technically challenging, especially for cities with extensive road networks. Recent advancements in machine learning and data-driven approaches have offered promising alternatives for estimating TMCs. Traffic patterns can vary significantly across different intersections due to factors such as road geometry, traffic signal settings, and local driver behaviors. This domain discrepancy limits the generalizability and accuracy of machine learning models when applied to new or unseen intersections. In response to these limitations, this research proposes a novel framework leveraging transfer learning (TL) to estimate TMCs at intersections by using traffic controller event-based data, road infrastructure data, and point-of-interest (POI) data. Evaluated on 30 intersections in Tucson, Arizona, the performance of the proposed TL model was compared with eight state-of-the-art regression models and achieved the lowest values in terms of Mean Absolute Error and Root Mean Square Error.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references

[1]

Adl, M., Ahmed, R., Vidal, C., & Emadi, A. (2024). Enhanced Vehicle Movement Counting at Intersections via a Self-Learning Fisheye Camera System. IEEE Access, 12(May), 77947–77958. https://doi.org/10.1109/ACCESS.2024.3408052 Bélisle, F., Saunier, N., Bilodeau, G. A., & Le Digabel, S. (2017). Optimized video tracking for automated vehicle turning movement ...

arXiv 2024
[261]

https://doi.org/10.1016/j.tra.2020.09.011 Ma, X., Karimpour, A., & Wu, Y. J. (2024). Data -driven transfer learning framework for estimating on-ramp and off-ramp traffic flows. Journal of Intelligent Transportation Systems: Technology, Planning, and Operations , 0(0), 1 –14. https://doi.org/10.1080/15472450.2023.2301696 Noh, H., Kramer, E., & Sun, A. (201...

arXiv 2024
[8369]

Yehia, A., Wang, X., Feng, M., Yang, X., & Gong, J. (2021). Applicability of boosting techniques in calibrating safety performance functions for freeways. Accident Analysis and Prevention, April, 106193. https://doi.org/10.1016/j.aap.2021.106193 Zhang, W., Zhu, F., Lv, Y., Tan, C., Liu, W., Zhang, X., & Wang, F. Y. (2022). AdapGL: An adaptive graph learni...

arXiv 2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Adl, M., Ahmed, R., Vidal, C., & Emadi, A. (2024). Enhanced Vehicle Movement Counting at Intersections via a Self-Learning Fisheye Camera System. IEEE Access, 12(May), 77947–77958. https://doi.org/10.1109/ACCESS.2024.3408052 Bélisle, F., Saunier, N., Bilodeau, G. A., & Le Digabel, S. (2017). Optimized video tracking for automated vehicle turning movement ...

arXiv 2024

[2] [261]

https://doi.org/10.1016/j.tra.2020.09.011 Ma, X., Karimpour, A., & Wu, Y. J. (2024). Data -driven transfer learning framework for estimating on-ramp and off-ramp traffic flows. Journal of Intelligent Transportation Systems: Technology, Planning, and Operations , 0(0), 1 –14. https://doi.org/10.1080/15472450.2023.2301696 Noh, H., Kramer, E., & Sun, A. (201...

arXiv 2024

[3] [8369]

Yehia, A., Wang, X., Feng, M., Yang, X., & Gong, J. (2021). Applicability of boosting techniques in calibrating safety performance functions for freeways. Accident Analysis and Prevention, April, 106193. https://doi.org/10.1016/j.aap.2021.106193 Zhang, W., Zhu, F., Lv, Y., Tan, C., Liu, W., Zhang, X., & Wang, F. Y. (2022). AdapGL: An adaptive graph learni...

arXiv 2021