Search

Forest fire is a significant issue, especially for tropical countries like Indonesia. One of the impacts of forest fires is environmental pollution and damage, such as damage to flora and fauna, water, and soil. Fire detection technology is crucial as a preventive measure before the spread or expansion of fire points. Several forest fire detection systems have been developed by various research studies, with detection targets varying. Objects in the form of images are usually detected using the RGB color filtering method, but this method still results in false detections in image processing. Therefore, a classification model is built to detect fire and smoke in images using the Convolutional Neural Network (ConvNet) algorithm. In the development of the ConvNet model, a comparison of models is also conducted to assess the influence of Hyperparameter Tuning and Callbacks in optimizing the model's classification performance. The research results indicate that out of the six comparison scenarios created, the best model is obtained with 90% training data and 10% testing data, which is also optimized with Hyperparameter Tuning and Callbacks, with a Validation Accuracy of 98.18% and Validation Loss of 4.97%. This model is then implemented in the interface system.

Offering promotion coupons is one of the most popular strategies of online marketing to attract new customers and increase customer loyalty. However, this strategy opens chances for fraud risk as the coupons are being redeemed multiple times using fake accounts. This risk becomes a burden to marketing costs and leads to failure to accomplish the intended strategic value. Therefore, this research focuses on building an automatic detection system of online promotion abuse based on its risk level. The proposed system also must work on live stream and bulk data. Therefore, in live stream data, it could alert the administrator before the transaction finished or the next process started. After conducting an exploratory factor analysis of the 24 attributes collected from four tables of data transaction, there were seven attributes indicating promotion abuse. These attributes were the user IP address, shipping address, mobile number, member email, order email, payment ID, and product name. Then, supervised machine learning of similarity algorithms was used to build models and find the hidden correlation of attributes to indicate the promotion abuse. The result from comparing five similarity methods showed that based on the workflow and performance, the most suitable methods for this case were exact match and Levenshtein edit base. The automatic risk scoring feature of the proposed system used seven attributes of online transactions as their most prominent promotion abuse parameter based on its hidden correlation. From the system performance testing, the result values of precision, recall, and F-measure are 95%, 93%, and 0.94, respectively. These results indicate that the system performance is satisfactory.

Strawberry is one type of fruit that is favored by the people of Indonesia. The detection process to identify strawberries can be done by utilizing advances in computer technology, One of them is in the field of digital image processing. In this study, we made a strawberry ripeness detection system using the values of Red, Green and Blue as the reference values, while for identification in determining the type of classification using the K-Means algorithm that uses the Euclidean distance difference as the reference. Based on the results of testing using the K-Means algorithm on 51 strawberry images consisting of ripe, semi ripe and raw fruit yielding an accuracy rate of 82.14%, we also conducted tests other than strawberry images as many as 8 images yielded an accuracy rate of 100%.