Implementation Of Unsupervised And Supervised Machine Learning To Determine Minimum Profit Margin In Agricultural Sector (CASE STUDY: CV RAWIN AGRO NUSA).

Dzaky, Achmadi Noor (2022) Implementation Of Unsupervised And Supervised Machine Learning To Determine Minimum Profit Margin In Agricultural Sector (CASE STUDY: CV RAWIN AGRO NUSA). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 024118400000067-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (10MB)

Abstract

Weather is the state of the atmosphere at a given time. Weather affects crop yields in the agricultural sector.CV Rawin Agro Nusa is contractual supplier to a catering company. The contract consists of the commodities required, and the length is 3 months. Therefore, this research study assesses the optimal strategy for CV Rawin Agro Nusa to generate sensible profit. A solution to this problem is proposed by generating a Machine Learning model to generate a prediction of future weather combined with heuristic aspects to determine the minimum profit margin that needs to be taken by CV Rawin Agro Nusa. The data collected from BMKG are the Average Temperature, Humidity, Sunlight Hours, and Precipitation. The data collected from CV Rawin Agro Nusa are the item profiles and the price changes matrix. The model used on this research are Logistic Regression, K-Means Clustering, and Random Forest Regression. Logistic Regression model is used to identify the existence of precipitation for a given period using the value of Humidity and Sunlight Hours in which it is able to reach an accuracy of 76%. The K-Means Clustering Model is used to identify the weather states from the weights or importances of each features in each clusters. There are four weather states generated from the K-Means Clustering. Random Forest model is used to predict the Average Temperature, Humidity and Sunlight Hours. Three Random Forest Regression models are used with different combinations of features. The best Random Forest model is Model 3 in which the model used only the Month and Date as an input. From the combination of all three models used, The prediction of the profit margin shows that for the period of January to March, the profit taken for Sensitive, Less-sensitive, and Non-sensitive items are 32%, 12%, and 3%. The accuracy of this prediction when compared to the average price of each item profiles are 94% for Sensitive items, 98% for Less-sensitive items, and 46% for Non-sensitive items.

Item Type:	Thesis (Other)
Additional Information:	RSI 658.403 55 Dza i-1 2022
Uncontrolled Keywords:	Weather Forecasting, Agriculture, Machine Learning, Profit Margin, Logistic Regression, K-Means Clustering, Random Forest.
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	12 May 2026 04:15
Last Modified:	12 May 2026 04:15
URI:	http://repository.its.ac.id/id/eprint/133157

Actions (login required)

View Item