[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121376-en":3,"doc-seo-121376-105":30,"detail-sidebar-cat-0-en-105":95},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},121376,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Machine Learning Approach for Site Classification - Thesis Summary","This thesis develops a machine learning approach for geotechnical site classification under GSFH-2, using MCCSTO descriptions covering moisture condition, colour, consistency, structure, texture, and origin. A flowchart was created with practicing engineers and geologists to classify text descriptions of 416 soil layers into the expected GSFH-2 categories. SVM, decision tree, and random forest models were trained using TF-IDF with NLP preprocessing such as lemmatisation, Porter stemming, and n-grams. Evaluation used feature importance, confusion matrices, and metrics including precision, recall, F1-score, and accuracy. The models achieved about 70.3% average accuracy, with the RF model reaching 71% on test data. Validation indicates that adding verified labelled data can improve accuracy and reduce overfitting to training data.","Machine Learning Approach for Site Classification  \nby  \nLaurence Armand Brits  \nThesis presented in partial fulfilment of the requirements for the degree of Master of Engineering (Civil) in the Faculty of Engineering at  \nStellenbosch University  \nSupervisor: Prof. C. MacRobert  \nMarch 2025  \nDeclaration  \nBy submitting this thesis electronically, I declare that the entirety of the work contained therein is my own, original work, that I am the sole author thereof (save to the extent explicitly otherwise stated), that reproduction and publication thereof by Stellenbosch University will not infringe any third party rights and that I have not previously in its entirety or in part submitted it for obtaining any qualification.  \n2024/12/01  \nDate:    \nCopyright © 2025 Stellenbosch University All rights reserved.  \nAbstract  \nMachine Learning Approach for Site Classification  \nL.A. Brits  \nDepartment of Civil Engineering,  \nStellenbosch University,  \nPrivate Bag X1, Matieland 7602, South Africa.  \nThesis: MEng (Civil)  \nMarch 2025  \nThis thesis presents a Machine Learning approach for site classification according to the Geotechnical Site Investigations for Housing Developments (GSFH-2) based on MCCSTO (moisture condition, colour, consistency, structure, texture and origin) descriptions. Based on input from practicing engineers and geologists a flowchart was developed to classify text descriptions of 416 individual soil layers into GSFH-2 classification. These soil layers were classified according to the expected soil movement by considering the descriptions of the moisture conditions, colour, consistency, structure, texture and origin of the soil. Three machine learning models, namely: Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF) models, were then developed using the database. Term Frequency–Inverse Document Frequency (TF-IDF) was used as an embedding technique in combination with other Natural Language Processing (NLP) methods, namely Lemmatisation, Porter-stemming, and N-grams to obtain the model that can best predict the classification of a given soil layer. To evaluate and analyse the results Feature Importance, Confusion Matrices, and statistical metrics (precision, recall, F-1 score, and accuracy) were used. The results revealed that the three models achieved an average accuracy of 70.3%, with the RF-model, utilizing only lower-casing as a preprocessing step, achieving the highest accuracy of 71% on the testing data. The RF-model was evaluated against a validation dataset and results indicated that the addition of verified labeled data can increase the accuracy of the proposed RF-model and reduce the overfitting of the model to the training data.  \nUittreksel  \nMasjienleer-benadering vir Terreinklassifikasie  \n(“Machine Learning Approach for Site Classification”)  \nL.A. Brits  \nDepartement Seviele Ingenieurswese,  \nStellenbosch Universiteit,  \nPrivaatsak X1, Matieland 7602, Suid Afrika.  \nTesis: MIng (Siviel)  \nMaart 2025  \nHierdie tesis bied ’n Masjienleer-benadering vir terreinklassifikasie volgens die Geotegniese Terreinondersoeke vir Behuisingsontwikkelings (GSFH-2) gebaseer op die MCCSTO-beskrywings. Op grond van insette van praktiserende ingenieurs en geoloë is ’n vloeidiagram ontwikkel om 416 individuele grondlaein GSFH-2-klassifikasie te kategoriseer. Hierdie grondlae is geklassifiseer volgens die verwagte grondbeweging deur die beskrywings van die vogtoestand, kleur, digtheid of styfheid, struktuur, tekstuur enoorsprong van die grond in ag te neem. Drie masjienleermodelle, naamlik: Ondersteuningsvektormasjien (SVM), Besluitboom (DT), en Willekeurige Woud (RF)-modelle, is met behulp van die databasis ontwikkel. Termfrekwensie–Omgekeerde Dokumentfrekwensie (TF-IDF) is as ’n inbeddingstegniek gebruik in kombinasie met ander NTP-metodes, soos lemmatisering, Porter-stamvorming en N-gramme, om die model te verkry wat die beste voorspellingsvermoë vir die klassifikasie van ’n gegewe grondlaag bi","cbCairQEIms6tcBJ","https://ap.wps.com/l/cbCairQEIms6tcBJ","pdf",7670008,1,119,"English","en",105,"# Introduction\n## Background\n## Research Scope\n## Objectives\n## Document Orginization\n# Literature Review\n## The Geotechnical Investigation\n### Desk Study and Walkover Survey\n### Intrusive Investigations\n### Site Classification\n## The Soil Profile\n### Basic Materials of the Soil Profile","[{\"question\":\"How does the thesis define input data for site classification under GSFH-2?\",\"answer\":\"It uses MCCSTO-based text descriptions that capture moisture condition, colour, consistency, structure, texture, and origin for soil layers. The approach converts these descriptions into GSFH-2 classes.\"},{\"question\":\"Which machine learning models are developed and compared?\",\"answer\":\"Three models are developed: Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF). The RF model is reported as the top performer on test data.\"},{\"question\":\"What method is used to convert text descriptions into model features?\",\"answer\":\"Term Frequency–Inverse Document Frequency (TF-IDF) is used as an embedding technique, combined with NLP steps such as lemmatisation, Porter stemming, and n-grams.\"},{\"question\":\"How are model performance and results evaluated?\",\"answer\":\"Evaluation relies on feature importance, confusion matrices, and statistical metrics including precision, recall, F-1 score, and accuracy. Validation further assesses how adding verified labelled data affects accuracy and overfitting.\"}]","Machine Learning Approach for Site Classification - Thesis Summary | PDF",1785735324,300,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":90,"head_meta":92,"extra_data":94,"updated_unix":28},"machine-learning-approach-for-site-classification-thesis-summary","",{"@graph":36,"@context":89},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-approach-for-site-classification-thesis-summary/121376/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"How does the thesis define input data for site classification under GSFH-2?","Question",{"text":75,"@type":76},"It uses MCCSTO-based text descriptions that capture moisture condition, colour, consistency, structure, texture, and origin for soil layers. The approach converts these descriptions into GSFH-2 classes.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which machine learning models are developed and compared?",{"text":80,"@type":76},"Three models are developed: Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF). The RF model is reported as the top performer on test data.",{"name":82,"@type":73,"acceptedAnswer":83},"What method is used to convert text descriptions into model features?",{"text":84,"@type":76},"Term Frequency–Inverse Document Frequency (TF-IDF) is used as an embedding technique, combined with NLP steps such as lemmatisation, Porter stemming, and n-grams.",{"name":86,"@type":73,"acceptedAnswer":87},"How are model performance and results evaluated?",{"text":88,"@type":76},"Evaluation relies on feature importance, confusion matrices, and statistical metrics including precision, recall, F-1 score, and accuracy. Validation further assesses how adding verified labelled data affects accuracy and overfitting.","https://schema.org",{"og:url":52,"og:type":91,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":93,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,132,135,139],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":128,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":130,"slug":131},9,"Religion & Spirituality",20,"religion-spirituality",{"id":130,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":130,"slug":134},"World Cup","world-cup",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":136,"slug":138},10,"Lifestyle","lifestyle",{"id":140,"doc_module":4,"doc_module_name":46,"category_name":141,"show_sort_weight":110,"slug":142},19,"General","general"]