The study area is one of the biggest gorges in southeast Ethiopia formed by the Wabe Shebelle river. The road wasconstructed to connect Gasera to Indeto. The study aimed to map and gives some mitigation strategies forgeohazards along the road corridor in Southeast Ethiopia using a GIS with AHP & FR technique within the 3 kmbuffer zones along the road corridor. The study included field investigations and inventorying, characterizationof geological situations, assessments of the hydrogeological conditions, and identification of slope instabilityvariables. A GIS technique was used to map an LSM with the combination of two models (AHP and FR). The studyused eight factors: slope, aspect, lithology, rainfall, altitude, road proximity, river proximity, and land use/cover.The results of the study revealed that LSZ maps performed using FR and AHP were 64.5 % and 69 % and-theinventory shows high and very high LSZ respectively. Rockfalls, debris/earth slides, and rockslides arecommonly observed landslides in the area. Based on the analysis lithology (basaltic and limestone formations)showed the highest contributions for landslide in the area. Slope and aspects show the most frequent landslidehazards in >40, 30–40◦, and east, and northeast respectively. Generally, the study found that lithology, slope,and aspect were the main factors contributing to slope instability in the study area. The produced landslidesusceptibility map is very important for urban planners, agricultural studies, environmentalists, and futurelandslide hazardous prevention and mitigation strategies.
Keywords: Landslides, LSM, AHP, FR, GIS, and Southeast Ethiopia

One of the biggest global challenges to development and public health is road traffic accidents (RTAs). As aresult, this study focuses on analysing road traffic accident determinant factors using the Wrapper Feature Se-lection Method in case of East Gojjam Zone located in Amhara region, Ethiopia, sub-Saharan. To do this, EastGojjam Road traffic office RTA data classified as simple injury, major injury, and death is gathered. The gatheredinformation is pre-processed before being used using machine learning classification algorithms includingNearest Neighbour (KNN), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), and NaïveBayes (NB). Using the wrapper feature selection approach, the most significant factor was identified using themachine-learning algorithm KNN, which obtained the best classification score with an accuracy of 99.5 %. Thus,the type of vehicle, the reason for the accident, the location of the accident, and the licence level were identifiedas crucial RTA factors. Finally, the variables, Sino track, unfavourable weather, Dolphin, and Debre Elias rated100 %, 100 %, 85 %, and 82.35 % for fatality in relation to the factors licence driver, cause of accident, type ofvehicle, and accident location, respectively.
Keywords: Road traffic accident, East Gojjam, Amhara region, Ethiopia, Machine learning, Feature selection, Filter, Wrapper method, Embedded method, Data mining

Now adays people express and share their opinions on various events on the internet thanks to socialmedia. Opinion mining is the process of interpreting user-generated opinion data on social media.Aside from its lack of resources in opinion-mining tasks, Amharic presents numerous difficultiesbecause of its complex structure and variety of dialects. Analyzing every comment written in Amharicis a challenging task. Significant advancements in opinion mining have been achieved using deeplearning. An opinion-mining model was used in this study to classify user comments written in Amharicas positive or negative. The domains that we focus on in this study are YouTube and Facebook. Fromthe Ethiopian broadcasts YouTube and Facebook official pages, we gathered 11,872 unstructured datafor this study using www.exportcomment.com, and Facebook page tools. Text preprocessing andfeature extraction techniques were used, in addition to manual annotation by linguistic specialists.The dataset was prepared for the experiment after annotation, preprocessing, and representation.LSTM, GRU, BiGRU, BiLSTM, and a hybrid of CNN with BiLSTM classifiers from the TensorFlow Kerasdeep learning library were used to train the model using the dataset, which was split using the 80/20train-test method, which proved effective for classification problems. Finally, we achieved of 94.27%,95.20%, 95.49%, 95.62%, and 96.08% using GRU, BiGRU, LSTM, BiLSTM, and CNN with BiLSTM,respectively, in word2vec embedding model.Keywords: Opinion mining, Deep learning, Recurrent neural network, Word2vec, Fast text

Natural fiber-reinforced polymer composites are the most widely used materials and preferable interms of biodegradability, cost production, recyclability, and low density. The main aim of thisstudy is to conduct an experimental investigation on tensile strength and impact strength ofpalmyra palm leaf stalk fiber (PLSF) and sisal fiber reinforced polymer hybrid composite. Thecomposite material was fabricated using hand lay-up techniques. The working parameters aremass fraction ratio of PLSF/sisal fiber and volume fiber fraction with the matrix. Tensile strengthand impact energy resistance tests were experimentally conducted according to the ASTM stan-dard dimensions. The results revealed that the addition of sisal fiber to PLSF enhanced the tensilestrength by 12.850 %, 26.540 %, and 30.630 % respectively compared to pure Palmyra palm leafstalk fiber reinforced composite (PPFRC). Whereas, the addition of PLSF to sisal fiber improvedthe impact of energy by 20.980 %, 13.610 %, and 11.880 % compared to pure sisal fiber rein-forced composite (PSFRC). The tensile strength with 20 % fiber volume fraction is improved by53.996 % and 12.188 % compared to 10 % and 15 % of fiber respectively. The impact strengthwas also enhanced by 24.931 % and 10.030 % compared to 10 % and 15 % of volume fiberfraction respectively. The tensile strength and impact energy of the treated fiber compositeincreased by 62.243 % and 22.478 % respectively compared to the untreated hybrid Palmyrapalm leaf stalk and sisal hybrid fiber reinforced composite (UHPSFRC). Generally, the HPSFRC-2(Palmyra palm leaf stalk/sisal fiber) (P/S ratio 50/50 % ratio with 20/80 % ratio of fiber/matricpercentage reinforced polymer hybrid composite) has good tensile strength and impact energy.Therefore, the mechanical property of the (Palm/Sisal) hybrid composite can be used for themanufacturing of the automotive interior parts like door panel, dash board, seat back, andautomotive roof.
Keywords: Handy lay-up, Hybrid fiber, Mechanical properties. unsaturated polyester resin

Every year, millions die in road accidents globally, imposing significant economic and humanitariancosts. While road traffic accidents are a major health concern, many developing countries, includingEthiopia, struggle to address this issue effectively. Ethiopia ranks second in East Africa for severe roadtraffic accidents, highlighting the need for improved injury reduction strategies. This study introduces anovel approach by chronologically identifying and prioritizing accident black spots in the studied area,Ethiopia. This method provides a valuable tool for transportation authorities and traffic police to targethigh-risk areas for immediate intervention. Focusing on the Dembecha-Injibara highway segment,the study employs both descriptive and inferential analyses, using the Zegeer method to calculateaccident rates. It also uses factors of weight contributing to road traffic accidents and their severityto rank accident-prone areas. The findings reveal that areas near Finote Selam, Banja, and Burie arehighly prone to severe accidents, with specific accident frequencies and priority values identified.Recommendations are offered to address these high-risk areas and mitigate severe traffic accidents inthe study region.Keywords: Road Traffic accidents, Severity, Prioritization and identification of Black Spot

Induction motor drives with variable speed applications that employ vector control are quitepopular nowadays because they provide strong dynamic performance and flexible speed control.By decoupling the torque-producing current components of stator current from the rotor flux,Indirect Field Oriented Control is recognized for generating excellent performance in inductionmotor drives. This investigation is being done to show the effectiveness of the novel FPI input-output data-trained ANFIS controller and compare the three controllers’ performance in termsof load variation capabilities, motor parameter variation, and speed tracking. Consequently, acomparison of the three controllers is important to select which controller performs high in in-duction motor drive. Indirect Field Oriented Control of induction motor with Fractional Pro-portional Integral (FPI), Integer Proportional Integral (IPI), and Adaptive Neuro-Fuzzy InferenceSystem (ANFIS) controllers are all discussed in this work along with their designs and compar-ative analysis. The square of error was used as a fitness function to genetically optimize the FPIand IPI controller parameters. The suggested Adaptive Neuro-Fuzzy Inference System (ANFIS)controller uses a hybrid learning approach. It is trained by the FPI controller’s input-output data.Using the results of MATLAB simulations under various operating situations, the performance ofthe ANFIS controller was compared with FPI and IPI controllers. Because of FPI controller in-cludes an extra parameter for adjustment, namely integration order, it performed better than IPIcontroller for speed control of the induction motor. According to the simulation findings, thepercentage peak overshoots while employing ANFIS, FPI, and IPI controllers were 0.495 %,12.062 %, and 14.699 % respectively. As a result, ANFIS exhibits a drastic reduction in overshoot.Additionally, with the ANFIS controlled induction motor drive, the speed achieves the requiredset value at 0.14 s. For no load, constant, and changing loads, the induction motor drive’s per-formance has been examined.
Keywords: Induction motor, Indirect field oriented control, Fractional PI, ANFIS, Integer PI

The groundwater level in the Chemoga watershed has been declining due to an increase in water demand, anthropogenicactivities, and climate change effects. This paper uses the WetSpass-MODFLOW coupling to evaluate the groundwater rechargein the Chemoga watershed. The MODFLOW groundwater flow simulation model is then used to simulate the hydraulic headdistribution based on these findings. The input data of WetSpass models are soil, land cover, topography, slope, and ground-water depth, as well as monthly meteorological characteristics (such as temperature, wind speed, and rainfall). The long-termspatial and temporal average annual precipitation of 1,453 mm is distributed as 169 mm (11.63%) groundwater recharge and879 mm (60.5%) surface runoff, while 405 mm (27.87%) is lost through evapotranspiration. In such seasonal variations, thegroundwater head due to the wet/summer stress period varied from 4 to 41 m. While in the dry/winter stress period ground-water head varied from 3.5 to 39.8 m, and also the groundwater head due to the annual stress period varied from 3.7 to 40 m.The findings are extensive and can be applied to water resource management and groundwater resource development in asustainable manner by safeguarding high groundwater recharge locations, and reevaluating allowable groundwater abstractionrates.Key words: ArcGis, Chemoga watershed, groundwater recharge, hydraulic head, MODFLOW, WetSpass-M model

Given the Amharic language has a lot of perplexing terminology since it features duplicate homophone letters,fidel’s ሀ, ሐ, and ኀ (three of which are pronounced as HA), ሠ and ሰ (both pronounced as SE), አ and ዐ (bothpronounced as AE), and ጸ and ፀ (both pronounced as TSE). The WSD (Word Sense Disambiguation) model, whichtackles the issue of lexical ambiguity in the context of the Amharic language, is developed using a deep learningtechnique. Due to the unavailability of the Amharic wordnet, a total of 1756 examples of paired Amharicambiguous homophonic words were collected. These words were ድህነት(dhnet) and ድኅነት(dhnet), ምሁር(m’hur)and ምሑር(m’hur), በአል(be’al) and በዢል(be’al), አቢይ (abiy) and ዐቢይ(abiy), with a total of 1756 examples.Following word preprocessing, word2vec, fasttext, Term Frequency-Inverse Document Frequency (TFIDF), andbag of words (BoW) were used to vectorize the text. The vectorized text was divided into train and test data. Thetrain data was then analysed using Naive Bayes (NB), K-nearest neighbour (KNN), logistic regression (LG), de-cision trees (DT), random forests (RF), and random oversampling technique. Bidirectional Gate Recurrent Unit(BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) improved to 99.99 % accuracy even with limiteddatasets.
Key Words: Amharic language, Homophone, Machine learning, Deep learning, Bidirectional, BiLSTM, BiGRU, TFIDF, BoW, Word embedding, Amharic word sense disambiguation

Landslides are prevalent in the Ethiopian highlands, particularly in the east Gojjam zone, which is highly affectedby landslide problems. This research was carried out in the east Gojjam zone, northwestern Ethiopia. The studyarea is part of an economically important area in the country, and it is the main source of water for the GrandEthiopian Renaissance Dam (GERD). The main objective of this work was to undertake a detailed inventory ofpast landslide locations and prediction of present and future landslide hazards, as well as the preparation of alandslide zonation map in the East Gojjam zone by using the Analytical Hierarchy Process (AHP) with the GIStechnique. The parameters used for this study were slope degree, slope aspect, land use and land cover, roadproximity, rainfall, lithology, altitude, and river proximity. The various causative parameters were collected fromthe field, and suitable modifications were made to the thematic maps. Finally, the ratings for various parameterswere used as the basis to prepare the LHZ map in GIS windows. The landslide susceptibility and inventorymapping were produced in the GIS environment. The results of the study show that the main driving factors forthe landslide hazards in the area were river proximity, rainfall, and manmade activities. Validation of this LHZmap revealed that more than 80% of past landslides match within the "high hazard zone" and reasonablyaccepted the rationality of the adopted methodology. The considered parameters, as well as their evaluation ofthe production of LHZ-Map, were confirmed. The produced landslide inventory map is very important for urbanplanners, agricultural studies, environmentalists, and future landslide hazardous prevention and mitigationstrategies.
Keywords: GIS, AHP, Inventory mapping, Causative factors, Landslides