pith. sign in

arxiv: 1907.11205 · v1 · pith:IY23WTR4new · submitted 2019-07-25 · 💻 cs.NI

Localization in Ultra Narrow Band IoT Networks: Design Guidelines and Trade-Offs

Hazem Sallouha , Alessandro Chiumento , Sreeraj Rajendran , Sofie Pollin This is my paper

Pith reviewed 2026-05-24 15:46 UTC · model grok-4.3

classification 💻 cs.NI
keywords IoT localizationRSSI fingerprintingultra narrowbandSigfoxmachine learning classificationGPS-enabled sensorsmultilaterationdevice-to-device
0
0 comments X

The pith

A handful of GPS sensors let machine learning classify IoT device locations at 80 percent accuracy using only RSSI in ultra-narrowband networks.

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

The paper shows how a few GPS-equipped sensors can divide a large coverage area into classes so that standard sensors without GPS can be located by matching their received signal strength patterns. Machine learning at the network center performs the classification, avoiding any extra hardware or power cost on the ordinary devices. Real measurements collected across Antwerp demonstrate that seven classes inside a 2.5 km radius city reach 80 percent accuracy, and widening the gaps between classes raises that figure to 92 percent. When the GPS sensors are dense enough for direct device links, multilateration further raises the chance of keeping localization error under 20 m by 40 percent.

Core claim

Splitting the service area into classes via a sparse set of GPS-enabled sensors (GSNs) produces repeatable RSSI fingerprints that standard classifiers can map to the correct class for ordinary sensors (SNs). This yields 80 percent classification accuracy for seven classes across a 2.5 km radius city and 92 percent when class spacing is increased; when GSN density permits device-to-device communication, multilateration improves the probability of sub-20 m error by 40 percent in the measured scenario.

What carries the argument

RSSI fingerprinting with machine-learning classifiers trained on signals received at a small number of GPS-enabled sensors to assign location classes to standard sensors.

If this is right

  • Classification accuracy reaches 80 percent when a 2.5 km radius city is divided into seven classes.
  • Increasing physical spacing between classes raises accuracy to 92 percent on the measured data.
  • Multilateration among densely placed GPS sensors improves the probability of localization error below 20 m by 40 percent.
  • The method requires no added payload, power, or hardware on the standard sensors.
  • Performance is validated on a real city-scale dataset collected in Antwerp.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same class-based fingerprinting could be tested in other long-range low-power networks to see whether similar accuracy holds without network-specific tuning.
  • Optimal class count and spacing might be derived analytically from city size and expected path-loss variation rather than chosen empirically.
  • Combining the RSSI classifier output with occasional coarse timing or angle information could further tighten the error distribution in sparse deployments.
  • If class boundaries are allowed to move over time, the approach might adapt to seasonal changes in propagation without retraining the entire model.

Load-bearing premise

The RSSI values collected from the GPS sensors produce sufficiently distinct and repeatable fingerprints within each class so that standard machine learning classifiers can reliably map new measurements to the correct class.

What would settle it

Collecting fresh RSSI traces at the same Antwerp locations and finding that the fingerprints no longer separate the seven classes at better than 70 percent accuracy would falsify the central performance claim.

Figures

Figures reproduced from arXiv: 1907.11205 by Alessandro Chiumento, Hazem Sallouha, Sofie Pollin, Sreeraj Rajendran.

Figure 1
Figure 1. Figure 1: Illustration of a long-range IoT network in which knowing at which [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A Histogram showing the variance of RSSI measurements at various [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Classifying SNs to one of the virtual GSN’s classes. For the case of RSSI, the CRLB of an estimated distance dˆ is given by the following inequality [10] q Var(dˆ) > ln 10 10 σsh np d , (2) where d is the distance between the node and the base station, np is the path loss exponent, and σsh represents the standard deviation of the log-normal shadowing effect. By detailing the variables in (2), we first have… view at source ↗
Figure 4
Figure 4. Figure 4: A map showing the nodes’ positions in the university campus with [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Maps showing the position of nodes divided into three, seven, and eighteen classes. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Classification accuracy versus the number of BSs (features), 7 classes, [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Classification accuracy versus the number of training messages from [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The confusion matrices for the case of three, seven, and eighteen adjacent classes with training data of 60, 40, and 20 messages from each [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Classification accuracy using RndF with different [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Classification accuracy of both SVM and RndF algorithms using the [PITH_FULL_IMAGE:figures/full_fig_p009_11.png] view at source ↗
Figure 10
Figure 10. Figure 10: The classification accuracy as a function of [PITH_FULL_IMAGE:figures/full_fig_p009_10.png] view at source ↗
Figure 12
Figure 12. Figure 12: Average RSSI values at three different receivers with a separation [PITH_FULL_IMAGE:figures/full_fig_p009_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Power regression and polynomial regression of RSSI measurements [PITH_FULL_IMAGE:figures/full_fig_p010_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Empirical CDF: performance of regression vs. fingerprinting for the [PITH_FULL_IMAGE:figures/full_fig_p010_14.png] view at source ↗
read the original abstract

Localization in long-range Internet of Things networks is a challenging task, mainly due to the long distances and low bandwidth used. Moreover, the cost, power, and size limitations restrict the integration of a GPS receiver in each device. In this work, we introduce a novel received signal strength indicator (RSSI) based localization solution for ultra narrow band (UNB) long-range IoT networks such as Sigfox. The essence of our approach is to leverage the existence of a few GPS-enabled sensors (GSNs) in the network to split the wide coverage into classes, enabling RSSI based fingerprinting of other sensors (SNs). By using machine learning algorithms at the network backed-end, the proposed approach does not impose extra power, payload, or hardware requirements. To comprehensively validate the performance of the proposed method, a measurement-based dataset that has been collected in the city of Antwerp is used. We show that a location classification accuracy of 80% is achieved by virtually splitting a city with a radius of 2.5 km into seven classes. Moreover, separating classes, by increasing the spacing between them, brings the classification accuracy up-to 92% based on our measurements. Furthermore, when the density of GSN nodes is high enough to enable device-to-device communication, using multilateration, we improve the probability of localizing SNs with an error lower than 20 m by 40% in our measurement scenario.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper proposes an RSSI-based fingerprinting approach for localization in ultra-narrowband IoT networks (e.g., Sigfox) that uses a small number of GPS-enabled sensors (GSNs) to partition coverage into classes; machine-learning classifiers at the backend map sensor-node (SN) RSSI vectors to these classes. On a measurement dataset collected in Antwerp, the method achieves 80% classification accuracy when virtually partitioning a 2.5 km radius city into seven classes; accuracy rises to 92% when class spacing is increased. When GSN density permits device-to-device links, multilateration is shown to raise the probability of localizing an SN with error below 20 m by 40%.

Significance. If the reported measurement results hold under proper validation, the work supplies concrete, deployment-relevant guidelines and trade-offs for RSSI fingerprinting in long-range UNB networks without imposing extra hardware, power, or payload costs on the majority of nodes. The Antwerp campaign and the explicit comparison of class-spacing versus multilateration regimes constitute the primary empirical contribution.

major comments (3)
  1. [Abstract, §4] Abstract and §4 (results): the headline claim that accuracy reaches 92% 'by increasing the spacing between them' is obtained via post-hoc adjustment of class boundaries. The manuscript must state whether class boundaries were chosen using only training data, whether the adjustment was performed once on the full dataset, and whether any form of nested cross-validation was used; otherwise the 92% figure risks optimistic bias.
  2. [Abstract, §3] Abstract and §3 (methodology): no error bars, confidence intervals, or details on the number of folds, random seeds, or hyper-parameter search are supplied for the 80% and 92% classification accuracies, nor for the 40% multilateration improvement. Standard practice requires at least k-fold cross-validation statistics and a description of the ML classifier family and its tuning procedure.
  3. [§5] §5 (multilateration scenario): the 40% improvement in P(error < 20 m) is conditioned on 'density of GSN nodes high enough to enable device-to-device communication.' The paper should quantify the minimum GSN density, the D2D range assumption, and the exact multilateration algorithm (e.g., least-squares, weighted) together with the number of participating GSNs per SN.
minor comments (2)
  1. [Figures 3-6] Figure captions and axis labels should explicitly state the number of GSNs used for each fingerprint and the train/test split ratio.
  2. [Abstract] The abstract states 'virtually splitting a city with a radius of 2.5 km into seven classes' without indicating how the class centroids or boundaries were initially placed; a short description or reference to the placement algorithm would improve reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which help strengthen the clarity and statistical rigor of the manuscript. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract, §4] Abstract and §4 (results): the headline claim that accuracy reaches 92% 'by increasing the spacing between them' is obtained via post-hoc adjustment of class boundaries. The manuscript must state whether class boundaries were chosen using only training data, whether the adjustment was performed once on the full dataset, and whether any form of nested cross-validation was used; otherwise the 92% figure risks optimistic bias.

    Authors: The 92% figure was produced by increasing class spacing on the full measurement dataset to illustrate the associated trade-off. We will revise the text to state explicitly that the adjustment used the complete dataset once, without nested cross-validation, and to note the exploratory character of this result. The core 80% accuracy result relies on a fixed, predefined partitioning and will be presented with the requested validation details. revision: yes

  2. Referee: [Abstract, §3] Abstract and §3 (methodology): no error bars, confidence intervals, or details on the number of folds, random seeds, or hyper-parameter search are supplied for the 80% and 92% classification accuracies, nor for the 40% multilateration improvement. Standard practice requires at least k-fold cross-validation statistics and a description of the ML classifier family and its tuning procedure.

    Authors: We will augment the manuscript with error bars and confidence intervals on all reported figures. The revised version will specify the cross-validation procedure (including the number of folds), the random seeds employed, the machine-learning classifier families, and the hyper-parameter search method used for both the classification and multilateration experiments. revision: yes

  3. Referee: [§5] §5 (multilateration scenario): the 40% improvement in P(error < 20 m) is conditioned on 'density of GSN nodes high enough to enable device-to-device communication.' The paper should quantify the minimum GSN density, the D2D range assumption, and the exact multilateration algorithm (e.g., least-squares, weighted) together with the number of participating GSNs per SN.

    Authors: We will expand §5 to report the minimum GSN density at which D2D links were observed in the Antwerp campaign, the D2D range assumption derived from the measurements, the multilateration algorithm (least-squares), and the average number of GSNs participating per SN in the evaluated scenario. revision: yes

Circularity Check

0 steps flagged

No significant circularity; results are empirical measurements

full rationale

The paper reports location classification accuracies (80-92%) and multilateration improvements (40%) directly from a measurement campaign on the Antwerp dataset using standard RSSI fingerprinting and machine learning classifiers. No equations, derivations, or fitted parameters are presented that reduce by construction to the inputs; the central claims rest on external data collection and off-the-shelf ML rather than self-referential definitions or self-citation chains. The approach is self-contained against the stated measurement benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The performance claims rest on the empirical separability of RSSI fingerprints in the chosen classes within the Antwerp measurement campaign and on the effectiveness of unspecified machine learning classifiers; no free parameters, mathematical axioms, or new physical entities are introduced.

axioms (1)
  • domain assumption RSSI measurements from GSNs yield class-distinct fingerprints that ML classifiers can exploit for SN localization
    Implicit premise required for the reported classification accuracies to hold

pith-pipeline@v0.9.0 · 5802 in / 1519 out tokens · 36652 ms · 2026-05-24T15:46:48.817202+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages

  1. [1]

    Localization in Long- range Ultra Narrow Band IoT Networks using RSSI,

    H. Sallouha, A. Chiumento, and S. Pollin, “Localization in Long- range Ultra Narrow Band IoT Networks using RSSI,” in 2017 IEEE International Conference on Communications (ICC) , pp. 1–6, 2017

    work page 2017

  2. [2]

    Sigfox and lorawan datasets for fingerprint localization in large urban and rural areas,

    M. Aernouts, R. Berkvens, K. Van Vlaenderen, and M. Weyn, “Sigfox and lorawan datasets for fingerprint localization in large urban and rural areas,” Data, vol. 3, no. 2, p. 13, 2018

    work page 2018

  3. [3]

    Internet of things for smart cities,

    A. Zanella, N. Bui, A. Castellani, L. Vangelista, and M. Zorzi, “Internet of things for smart cities,” IEEE Internet of Things journal, vol. 1, no. 1, pp. 22–32, 2014

    work page 2014

  4. [4]

    Location of things (lot): A review and taxonomy of sensors localization in iot infrastructure,

    R. C. Shit, S. Sharma, D. Puthal, and A. Y . Zomaya, “Location of things (lot): A review and taxonomy of sensors localization in iot infrastructure,” IEEE Communications Surveys & Tutorials , vol. 20, no. 3, pp. 2028–2061, 2018

    work page 2028

  5. [5]

    ulora: Ultra low-power, low-cost and open platform for the lora networks,

    H. Sallouha, B. Van den Bergh, Q. Wang, and S. Pollin, “ulora: Ultra low-power, low-cost and open platform for the lora networks,” in 4th ACM Workshop HotWireless ’17, pp. 43–47, 2017

    work page 2017

  6. [6]

    Sigfox: http://www.sigfox.com, [online] Accessed on: Jan. 2019

    work page 2019

  7. [7]

    Weightless-P: http://www.weightless.org/, [online] Accessed on: Jan. 2019

    work page 2019

  8. [8]

    alliance: http://lora-alliance.org, [online] Accessed on: Jan

    L. alliance: http://lora-alliance.org, [online] Accessed on: Jan. 2019

    work page 2019

  9. [9]

    NB-IoT: https://goo.gl/DaUHkV, [online] Accessed on: Jan. 2019

    work page 2019

  10. [10]

    Localization via ultra-wideband radios: a look at positioning aspects for future sensor networks,

    S. Gezici et al. , “Localization via ultra-wideband radios: a look at positioning aspects for future sensor networks,” IEEE signal processing magazine, vol. 22, no. 4, pp. 70–84, 2005

    work page 2005

  11. [11]

    Wireless rssi fingerprinting localization,

    S. Yiu et al., “Wireless rssi fingerprinting localization,” Signal Process- ing, vol. 131, pp. 235–244, 2017

    work page 2017

  12. [12]

    Base station ordering for emergency call localization in ultra-dense cellular networks,

    H. Elsawy, W. Dai, M.-S. Alouini, and M. Z. Win, “Base station ordering for emergency call localization in ultra-dense cellular networks,” IEEE Access, vol. 6, pp. 301–315, 2018

    work page 2018

  13. [13]

    A theoretical foundation of network localization and navigation,

    M. Z. Win, Y . Shen, and W. Dai, “A theoretical foundation of network localization and navigation,” Proceedings of the IEEE , vol. 106, no. 7, pp. 1136–1165, 2018

    work page 2018

  14. [14]

    Mercury: An infrastructure-free system for network localization and navigation,

    Z. Liu, W. Dai, and M. Z. Win, “Mercury: An infrastructure-free system for network localization and navigation,” IEEE Transactions on Mobile Computing, vol. 17, no. 5, pp. 1119–1133, 2018

    work page 2018

  15. [15]

    Network localization and navigation via cooperation,

    M. Z. Win, A. Conti, S. Mazuelas, Y . Shen, W. M. Gifford, D. Dardari, and M. Chiani, “Network localization and navigation via cooperation,” IEEE Communications Magazine , vol. 49, no. 5, pp. 56–62, 2011

    work page 2011

  16. [16]

    Network op- eration strategies for efficient localization and navigation,

    M. Z. Win, W. Dai, Y . Shen, G. Chrisikos, and H. V . Poor, “Network op- eration strategies for efficient localization and navigation,” Proceedings of the IEEE , vol. 106, no. 7, pp. 1224–1254, 2018. 11

    work page 2018

  17. [17]

    Efficient multisensor localization for the internet of things: Exploring a new class of scalable localization algorithms,

    M. Z. Win, F. Meyer, Z. Liu, W. Dai, S. Bartoletti, and A. Conti, “Efficient multisensor localization for the internet of things: Exploring a new class of scalable localization algorithms,” IEEE Signal Processing Magazine, vol. 35, no. 5, pp. 153–167, 2018

    work page 2018

  18. [18]

    Nextme: Localization using cellular traces in internet of things.,

    D. Zhang, S. Zhao, L. T. Yang, M. Chen, Y . Wang, and H. Liu, “Nextme: Localization using cellular traces in internet of things.,” IEEE Trans. Industrial Informatics, vol. 11, no. 2, pp. 302–312, 2015

    work page 2015

  19. [19]

    Indoor positioning: A key enabling technology for iot applications,

    D. Macagnano, G. Destino, and G. Abreu, “Indoor positioning: A key enabling technology for iot applications,” in Internet of Things (WF- IoT), 2014 IEEE World Forum on , pp. 117–118, IEEE, 2014

    work page 2014

  20. [20]

    Location of things (lot): A review and taxonomy of sensors localization in iot infrastructure,

    R. C. Shit, S. Sharma, D. Puthal, and A. Y . Zomaya, “Location of things (lot): A review and taxonomy of sensors localization in iot infrastructure,” IEEE Communications Surveys & Tutorials , 2018

    work page 2018

  21. [21]

    An energy-efficient and lightweight indoor localization system for internet-of-things (iot) environments,

    M. Kwak et al., “An energy-efficient and lightweight indoor localization system for internet-of-things (iot) environments,” ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies , vol. 2, no. 1, p. 17, 2018

    work page 2018

  22. [22]

    Path-loss-based fingerprint lo- calization approach for location-based services in indoor environments,

    J. Zhang, G. Han, N. Sun, and L. Shu, “Path-loss-based fingerprint lo- calization approach for location-based services in indoor environments,” IEEE Access, vol. 5, pp. 13756–13769, 2017

    work page 2017

  23. [23]

    Indoor localization for iot applica- tions using fingerprinting,

    P. Fonseka and K. Sandrasegaran, “Indoor localization for iot applica- tions using fingerprinting,” in IEEE World Forum on Internet of Things, WF-IoT 2018, 2018

    work page 2018

  24. [24]

    Csi amplitude fingerprinting-based nb-iot indoor local- ization,

    Q. Song et al., “Csi amplitude fingerprinting-based nb-iot indoor local- ization,” IEEE Internet of Things Journal , vol. 5, no. 3, 2018

    work page 2018

  25. [25]

    A comparative study on machine learning algorithms for indoor positioning,

    S. Bozkurt et al., “A comparative study on machine learning algorithms for indoor positioning,” in International Symposium on Innovations in Intelligent SysTems and Applications (INISTA) , pp. 1–8, IEEE, 2015

    work page 2015

  26. [26]

    Support vector machines for indoor sensor localization,

    W. Farjow, A. Chehri, M. Hussein, and X. Fernando, “Support vector machines for indoor sensor localization,” in 2011 IEEE Wireless Com- munications and Networking Conference , pp. 779–783, IEEE, 2011

    work page 2011

  27. [27]

    Gaussian process regression for fingerprinting based localization,

    S. Kumar et al. , “Gaussian process regression for fingerprinting based localization,” Ad Hoc Networks , vol. 51, pp. 1–10, 2016

    work page 2016

  28. [28]

    Ubiquitous localization (ubiloc): A survey and taxon- omy on device free localization for smart world,

    R. C. Shit et al., “Ubiquitous localization (ubiloc): A survey and taxon- omy on device free localization for smart world,” IEEE Communications Surveys & Tutorials, 2019

    work page 2019

  29. [29]

    Probabilistic rss fingerprinting for localization in smart platforms,

    R. C. Shit, S. Sharma, D. Puthal, and S. S. Tripathi, “Probabilistic rss fingerprinting for localization in smart platforms,” in 2018 International Conference on Information Technology (ICIT) , pp. 254–259, IEEE, 2018

    work page 2018

  30. [30]

    Device-free presence detection and localization with svm and csi fingerprinting,

    R. Zhou, X. Lu, P. Zhao, and J. Chen, “Device-free presence detection and localization with svm and csi fingerprinting,”IEEE Sensors Journal, vol. 17, no. 23, pp. 7990–7999, 2017

    work page 2017

  31. [31]

    Localization in wireless sensor networks based on support vector machines,

    D. A. Tran and T. Nguyen, “Localization in wireless sensor networks based on support vector machines,” IEEE Transactions on Parallel and Distributed Systems, vol. 19, no. 7, pp. 981–994, 2008

    work page 2008

  32. [32]

    Towards location enhanced iot: Characterization of lora signal for wide area localization,

    Y . Liy, Z. He, Y . Li, H. Xu, L. Pei, and Y . Zhang, “Towards location enhanced iot: Characterization of lora signal for wide area localization,” in 2018 Ubiquitous Positioning, Indoor Navigation and Location-Based Services (UPINLBS), pp. 1–7, IEEE, 2018

    work page 2018

  33. [33]

    Robust localization for cognitive iot via the mobile anchor node based on the diameter-varying spiral line,

    X. Wang, Z. Qian, X. Wang, and L. Huang, “Robust localization for cognitive iot via the mobile anchor node based on the diameter-varying spiral line,” IEEE Access, vol. 7, pp. 28487–28497, 2019

    work page 2019

  34. [34]

    Energy-constrained uav trajectory design for ground node localization,

    H. Sallouha, M. M. Azari, and S. Pollin, “Energy-constrained uav trajectory design for ground node localization,” in 2018 IEEE Global Communications Conference (GLOBECOM) , pp. 1–7, IEEE, 2018

    work page 2018

  35. [35]

    Localization in low power wide area networks using wi-fi fingerprints,

    T. Janssen, M. Weyn, and R. Berkvens, “Localization in low power wide area networks using wi-fi fingerprints,” Applied Sciences, vol. 7, no. 9, p. 936, 2017

    work page 2017

  36. [36]

    A comparison of signal strength localization methods with sigfox,

    M. Aernouts et al. , “A comparison of signal strength localization methods with sigfox,” in 15th WPNC, pp. 1–6, IEEE, 2018

    work page 2018

  37. [37]

    TD: https://github.com/Telecom-Design/Documentation TD, [online] Accessed on: Jan. 2019

    work page 2019

  38. [38]

    Range and coexistence analysis of long range unlicensed communication,

    B. Reynders, W. Meert, and S. Pollin, “Range and coexistence analysis of long range unlicensed communication,” in Telecommunications (ICT), 2016 23rd International Conference on , pp. 1–6, IEEE, 2016

    work page 2016

  39. [39]

    Scikit-learn: Machine learning in Python,

    F. Pedregosa et al., “Scikit-learn: Machine learning in Python,” Journal of Machine Learning Research , vol. 12, pp. 2825–2830, 2011

    work page 2011

  40. [40]

    Performance com- parison of support vector machine, random forest, and extreme learning machine for intrusion detection,

    I. Ahmad, M. Basheri, M. Iqbal, and A. Raheem, “Performance com- parison of support vector machine, random forest, and extreme learning machine for intrusion detection,” IEEE Access, 2018

    work page 2018

  41. [41]

    Gaussian process regression for csi and feedback estimation in lte,

    A. Chiumento et al., “Gaussian process regression for csi and feedback estimation in lte,” in 2015 IEEE International Conference on Commu- nication Workshop (ICCW), pp. 1440–1445, June 2015

    work page 2015

  42. [42]

    Adaptive CSI and feedback estimation in LTE and beyond: a Gaussian process regression approach,

    A. Chiumento et al., “Adaptive CSI and feedback estimation in LTE and beyond: a Gaussian process regression approach,” EURASIP Journal on Wireless Communications and Networking , vol. 2015, p. 168, Jun 201

    work page 2015

  43. [43]

    Indoor localization using improved rss-based lateration methods,

    J. Yang and Y . Chen, “Indoor localization using improved rss-based lateration methods,” in Global Telecommunications Conference, 2009. GLOBECOM 2009. IEEE , pp. 1–6, IEEE, 2009

    work page 2009

  44. [44]

    Preventing over-fitting during model selection via bayesian regularisation of the hyper-parameters,

    G. C. Cawley and N. L. Talbot, “Preventing over-fitting during model selection via bayesian regularisation of the hyper-parameters,” Journal of Machine Learning Research , vol. 8, no. Apr, pp. 841–861, 2007

    work page 2007