Machine Learning-based identification of Petroleum Distillates and Gasoline traces using measured and synthetic GC spectra from collected samples

Omer Kaspi, System Engineering, Afeka College of Engineering, Rosh Haain, Israel (omerkaspi@gmail.com)

Cases involving arsons are typically handled by forensic experts who examine spectra of samples collected from scenes of fire to test for the existence or absence of ignitable liquids. This is tedious work, since many cases do not involve such liquids. To facilitate this process, we have developed in this work a Machine Learning (ML)-based workflow for samples’ classification based on their GC chromatograms (i.e., spectra). To this end, annotated spectra of 181 samples containing three groups of liquids (Petroleum Distillates, Gasoline and an assortment of other substances) collected from fire scenes as well as two reference databases were obtained from the Israeli Department of Identification and Forensic Sciences (DIFS). These spectra were used for the derivation of ML-based classification models using three algorithms, namely, kNN, Representative Spectrum, and Random Forest (RF) giving rise to reliable predictions. To increase the size of the dataset to a level that would enable the usage of more advanced ML algorithms, we have used the experimental spectra to develop a new spectra synthesis algorithm and utilized it to generate a large dataset of synthetic spectra. This dataset was used for the derivation of new kNN, RF and Representative Spectrum models as well as Deep Learning (DL) models producing F1-scores over an independent test set composed entirely of “real” spectra ranging from 0.74-0.95, 0.86-0.95, 0.55-0.75, and 0.85-0.96 for kNN, RF, Representative Spectrum, and DL, respectively. Following the completion of the work, a second set of real spectra was provided to us by DIFS and modeling it with the second set of models yielded F1-scores ranging from 0.92-0.96, 0.96-1.00, 0.71-0.82, and 0.95-0.98 for kNN, RF, Representative Spectrum, and DL, respectively. These results therefore suggest that for this dataset, performances depend more on the size of the dataset used for model training than on the ML algorithm.

Short Biography of Presenting Author

As a seasoned professional in Systems Engineering and AI, I bring over a decade of experience in the defense industry, specializing in COMINT, SIGINT, and COMJAM systems. My expertise encompasses providing comprehensive solutions that integrate components from various disciplines, including Command and Control (C2), EO/IR, and Acoustics.

I founded Kaspi Solutions, a systems engineering-oriented advisory company, where I guide product development and manage multiple teams of up to twenty skilled professionals. Prior to establishing Kaspi Solutions, I managed the C-UAS Engineering department at Elbit Systems, leading a team of 11 experts and overseeing the R&D process for the department.

While my career has primarily focused on Systems Engineering, I have also led algorithm design and data science initiatives. I hold a Ph.D. in AI Applications in Forensics from Bar-Ilan University, with my research published in multiple peer-reviewed journals and presented at international conferences. My work particularly emphasizes machine learning applications in forensics and materials informatics .

I am passionate about leveraging cutting-edge technologies to tackle complex challenges and drive innovation in both academic and industrial settings.


Machine Learning-based identification of Petroleum Distillates and Gasoline traces using measured and synthetic GC spectra from collected samples Omer Kaspi, System Engineering, Afeka College of Engineering, Rosh Haain, Israel (omerkaspi@gmail.com) Cases involving arsons are typically handled by forensic experts who examine spectra of samples collected from scenes of fire to test for the existence or absence of ignitable liquids. This is tedious work, since many cases do not involve such liquids. To facilitate this process, we have developed in this work a Machine Learning (ML)-based workflow for samples’ classification based on their GC chromatograms (i.e., spectra). To this end, annotated spectra of 181 samples containing three groups of liquids (Petroleum Distillates, Gasoline and an assortment of other substances) collected from fire scenes as well as two reference databases were obtained from the Israeli Department of Identification and Forensic Sciences (DIFS). These spectra were used for the derivation of ML-based classification models using three algorithms, namely, kNN, Representative Spectrum, and Random Forest (RF) giving rise to reliable predictions. To increase the size of the dataset to a level that would enable the usage of more advanced ML algorithms, we have used the experimental spectra to develop a new spectra synthesis algorithm and utilized it to generate a large dataset of synthetic spectra. This dataset was used for the derivation of new kNN, RF and Representative Spectrum models as well as Deep Learning (DL) models producing F1-scores over an independent test set composed entirely of “real” spectra ranging from 0.74-0.95, 0.86-0.95, 0.55-0.75, and 0.85-0.96 for kNN, RF, Representative Spectrum, and DL, respectively. Following the completion of the work, a second set of real spectra was provided to us by DIFS and modeling it with the second set of models yielded F1-scores ranging from 0.92-0.96, 0.96-1.00, 0.71-0.82, and 0.95-0.98 for kNN, RF, Representative Spectrum, and DL, respectively. These results therefore suggest that for this dataset, performances depend more on the size of the dataset used for model training than on the ML algorithm. Short Biography of Presenting Author As a seasoned professional in Systems Engineering and AI, I bring over a decade of experience in the defense industry, specializing in COMINT, SIGINT, and COMJAM systems. My expertise encompasses providing comprehensive solutions that integrate components from various disciplines, including Command and Control (C2), EO/IR, and Acoustics. I founded Kaspi Solutions, a systems engineering-oriented advisory company, where I guide product development and manage multiple teams of up to twenty skilled professionals. Prior to establishing Kaspi Solutions, I managed the C-UAS Engineering department at Elbit Systems, leading a team of 11 experts and overseeing the R&D process for the department. While my career has primarily focused on Systems Engineering, I have also led algorithm design and data science initiatives. I hold a Ph.D. in AI Applications in Forensics from Bar-Ilan University, with my research published in multiple peer-reviewed journals and presented at international conferences. My work particularly emphasizes machine learning applications in forensics and materials informatics . I am passionate about leveraging cutting-edge technologies to tackle complex challenges and drive innovation in both academic and industrial settings.

Organized & Produced by:	www.isranalytica.org.il POB 4043, Ness Ziona 70400, Israel Tel.: +972-8-9313070, Fax: +972-8-9313071 Site: www.bioforum.org.il, E-mail: hagit@bioforum.co.il