Abstract: Prediction of solar power involves the knowledge of the sun , atmosphere and other parameters, and the scattering processes and the specifications of a solar energy plant that employs the sun's energy to generate solar power . This prediction result is essential for an efficient use of the solar power plant, the management of the electricity grid, and solar energy trading. However, because of nonlinear and nonstationary behavior of solar power time series, an efficient forecasting model is needed to predict it. Accordingly, in this paper, we propose a new forecast approach based on combination of a neural network with a metaheuristic algorithm as the hybrid forecasting engine. The metaheuristic algorithm optimizes the free parameters of the neural network. This approach also includes a 2‐stage feature selection filter based on the information‐theoretic criteria of mutual information and interaction gain, which filters out the ineffective input features. To demonstrate the effectiveness of the proposed forecast approach, it is implemented on a real‐world engineering test case. Obtained results illustrate the superiority of the proposed approach in comparison with other prediction methods.

Abstract: Cyber threats are a showstopper for Internet of Things (IoT) has recently been used at an industrial scale. Network layer attacks on IoT can cause significant disruptions and loss of information. Among such attacks, routing attacks are especially hard to defend against because of the ad-hoc nature of IoT systems and resource constraints of IoT devices. Hence, an efficient approach for detecting and predicting IoT attacks is needed. Systems confidentiality, integrity and availability depends on continuous security and robustness against routing attacks. We propose a deep-learning based machine learning method for detection of routing attacks for IoT. In our study, the Cooja IoT simulator has been utilized for generation of high-fidelity attack data, within IoT networks ranging from 10 to 1000 nodes. We propose a highly scalable, deep-learning based attack detection methodology for detection of IoT routing attacks which are decreased rank, hello-flood and version number modification attacks, with high accuracy and precision. Application of deep learning for cyber-security in IoT requires the availability of substantial IoT attack data and we believe that the IoT attack dataset produced in this work can be utilized for further research.

Abstract: The first successful isolation‐based anomaly detector, ie, iForest, uses trees as a means to perform isolation. Although it has been shown to have advantages over existing anomaly detectors, we have identified 4 weaknesses, ie, its inability to detect local anomalies, anomalies with a high percentage of irrelevant attributes, anomalies that are masked by axis‐parallel clusters, and anomalies in multimodal data sets.

Abstract: This paper describes an approach that combines generative adversarial networks (GANs) with interactive evolutionary computation (IEC). While GANs can be trained to produce lifelike images, they are normally sampled randomly from the learned distribution, providing limited control over the resulting output. On the other hand, interactive evolution has shown promise in creating various artifacts such as images, music and 3D objects, but traditionally relies on a hand-designed evolvable representation of the target domain. The main insight in this paper is that a GAN trained on a specific target domain can act as a compact and robust genotype-to-phenotype mapping (i.e. most produced phenotypes do resemble valid domain artifacts). Once such a GAN is trained, the latent vector given as input to the GAN’s generator network can be put under evolutionary control, allowing controllable and high-quality image generation. In this paper, we demonstrate the advantage of this novel approach through a user study in which participants were able to evolve images that strongly resemble specific target images.

Abstract: This paper proposes a novel methodology for the detection of intrusive comments or spam on the video-sharing website - Youtube. We describe spam comments as those which have a promotional intent or those who deem to be contextually irrelevant for a given video. The prospects of monetisation through advertising on popular social media channels over the years has attracted an increasingly larger number of users. This has in turn led to to the growth of malicious users who have begun to develop automated bots, capable of large-scale orchestrated deployment of spam messages across multiple channels simultaneously. The presence of these comments significantly hurts the reputation of a channel and also the experience of normal users. Youtube themselves have tackled this issue with very limited methods which revolve around blocking comments that contain links. Such methods have proven to be extremely ineffective as Spammers have found ways to bypass such heuristics. Standard machine learning classification algorithms have proven to be somewhat effective but there is still room for better accuracy with new approaches. In this work, we attempt to detect such comments by applying conventional machine learning algorithms such as Random Forest, Support Vector Machine, Naive Bayes along with certain custom heuristics such as N-Grams which have proven to be very effective in detecting and subsequently combating spam comments.

Abstract: In modern society, Heart disease is the noteworthy reason for short life. Large population of people depends on the healthcare system so that they can get accurate result in less time. Large amount of data is produced and collected by the healthcare organization on the daily basis. To get intriguing knowledge, data innovation permits to extract the data through automization of processes. Weighted Association Rule is a type of data mining technique used to eliminate the manual task which also helps in extracting the data directly from the electronic records. This will help in decreasing the cost of services and also helps in saving lives. In this paper, we will find the rule to predict patient's risk of having coronary disease. Test results have shown that vast majority of the rules helps in the best prediction of coronary illness.

Abstract: With the advent of Deep Learning,the machine learning systems are able to recognize and classify objects of interest in an image.Various advancement has been done in the field of object recognition and classification.Our research work focusses on improving the R-CNN, Fast R-CNN,YOLO architecture.The work focussed on using Region Proposals Network(RPN) to extract region of interest in an image.RPN outputs an image based on the objectness score.The output objects are subjected to Roll Polling for classification.Our research work focusses on training Faster R-CNN using custom based data set of images. Our trained network efficiently detects objects from an image consisting of multiple objects.Our network requires minimum GPU capability of 3.0 or higher.

Abstract: Anomaly detection in streaming data is an important problem in numerous application domains. Most existing model‐based approaches to stream learning are based on decision trees due to their fast construction speed. This paper introduces streaming autoencoder (SA), a fast and novel anomaly detection algorithm based on ensembles of neural networks for evolving data streams. It is a one‐class learner, which only requires data from the positive class for training and is accurate even when anomalous training data are rare. It features an ensemble of threaded autoencoders with continuous learning capacity. Furthermore, the SA uses a 2‐step detection mechanism to ensure that real anomalies are detected with low false‐positive rates. The method is highly efficient because it processes data streams in parallel with multithreads and alternating buffers. Our analysis shows that SA has a linear runtime and requires constant memory space. Empirical comparisons to the state‐of‐the‐art methods on multiple benchmark data sets demonstrate that the proposed method detects anomalies efficiently with fewer false alarms.

Abstract: With the increase of volume, velocity, and variety of big data, the traditional collaborative filtering recommendation algorithm, which recommends the items based on the ratings from those like-minded users, becomes more and more inefficient. In this paper, two varieties of algorithms for collaborative filtering recommendation system are proposed. The first one uses the improved k-means clustering technique while the second one uses the improved k-means clustering technique coupled with Principal Component Analysis as a dimensionality reduction method to enhance the recommendation accuracy for big data. The experimental results show that the proposed algorithms have better recommendation performance than the traditional collaborative filtering recommendation algorithm.

Abstract: The purpose of this study is to use the truncated Newton method in prior correction logistic regression (LR). A regularization term is added to prior correction LR to improve its performance, which results in the truncated-regularized prior correction algorithm. The performance of this algorithm is compared with that of weighted LR and the regular LR methods for large imbalanced binary class data sets. The results, based on the KDD99 intrusion detection data set, and 6 other data sets at both the prior correction and the weighted LRs have the same computational efficiency when the truncated Newton method is used in both of them. A higher discriminative performance, however, resulted from weighting, which exceeded both the prior correction and the regular LR on nearly all the data sets. From this study, we conclude that weighting outperforms both the regular and prior correction LR models in most data sets and it is the method of choice when LR is used to evaluate imbalanced and rare event data.

Abstract: Swarm intelligence is one of the most promising area of numerical optimization to solve real‐world optimization problems. Grey wolf optimizer (GWO), which is based on leadership hierarchy of grey wolves, is one of the relatively new algorithm in the field of swarm intelligence–based algorithms. In order to solve constrained real‐world optimization problems, in this paper, a constrained version of GWO has been proposed by incorporating a simple constraint handling technique in GWO, and then an attempt is made to improve the ability of the leaders in original GWO by proposing random walk GWO (RW‐GWO) by pointing out some drawbacks in their process of searching prey. (To the best of the knowledge of the authors, a constrained version of GWO has not been developed yet. The unconstrained version of RW‐GWO has been proposed in the authors' earlier work.) The efficiency of both these proposed algorithms have been tested on the Institute of Electrical and Electronics Engineers Congress on Evolutionary Computation 2006 benchmark problems and on 3 engineering application problems to observe their comparative performance. It is concluded from the results that the proposed improved version of GWO, namely, RW‐GWO, has better potential to solve these constraint problems compared to GWO very efficiently as a constrained optimizer.

Abstract: Chatbots are computer programs capable to carry a conversation with human. They can be seen as an artificial agent designed to serve the purpose of conversation with the end user. Chatbots are gaining popularity especially in business and health sector as they have the potential to automate service and reduce human efforts. Widespread use of Apps, maturation of Artificial Intelligence (AI) technologies and integration of Natural Language Processing (NLP) fuels up the growth of chatbot. In this paper, we present different models of chatbots along with an architectural overview of computationally intelligent chatbot in context of recent technologies. In the three layer architecture, we have given insights of how the NLP, Natural Language Understanding (NLU) and Decision engine work together with Knowledge Base to achieve AI using Recurrent Neural Network (RNN) and Long Short Term Memory (LSTM). In addition, we also discuss different chatbot platforms and development frameworks of recent times. Our core emphasis is on analysis of recent development approaches of textbased conversational systems. We identify few challenges in intelligent chatbot development that may be helpful for future research works.

Abstract: One of the problems with traditional genetic algorithms (GAs) is premature convergence, which makes them incapable of finding good solutions to the problem. The memetic algorithm (MA) is an extension of the GA. It uses a local search method to either accelerate the discovery of good solutions, for which evolution alone would take too long to discover, or reach solutions that would otherwise be unreachable by evolution or a local search method alone. In this paper, we introduce a new algorithm based on learning automata (LAs) and an MA, and we refer to it as LA‐MA. This algorithm is composed of 2 parts: a genetic section and a memetic section. Evolution is performed in the genetic section, and local search is performed in the memetic section. The basic idea of LA‐MA is to use LAs during the process of searching for solutions in order to create a balance between exploration performed by evolution and exploitation performed by local search. For this purpose, we present a criterion for the estimation of success of the local search at each generation. This criterion is used to calculate the probability of applying the local search to each chromosome. We show that in practice, the proposed probabilistic measure can be estimated reliably. On the basis of the relationship between the genetic section and the memetic section, 3 versions of LA‐MA are introduced. LLA‐MA behaves according to the Lamarckian learning model, BLA‐MA behaves according to the Baldwinian learning model, and HLA‐MA behaves according to both the Baldwinian and Lamarckian learning models. To evaluate the efficiency of these algorithms, they have been used to solve the graph isomorphism problem. The results of computer experimentations have shown that all the proposed algorithms outperform the existing algorithms in terms of quality of solution and rate of convergence.

Abstract: This research aims to present a general framework by which the most appropriate wavelet parameters including mother wavelet, vanishing moment, and decomposition level can be chosen for a joint wavelet transform and machine learning model. This study is organized in 2 parts: the first part presents an evolutionary Levenberg-Marquardt neural network (ELMNN) model as the most effective machine learning configuration, and the second part describes how the wavelet transform can be effectively embedded with the developed ELMNN model. In this research, the rainfall and runoff time series data of 2 distinct watersheds at 2 different time scales (daily and monthly) were used to build the proposed hybrid wavelet transform and ELMNN model. The conclusions of this study showed that the Daubechies wavelet more than other wavelet families is capable to extract the informative features of hydrologic series. The vanishing moment and decomposition level of this mother wavelet should be selected based on the watershed behavior and the time resolution of rainfall and runoff time series, respectively. The verification results for both watersheds at daily and monthly time scales indicated root mean square error, peak value criterion, low value criterion, and Kling-Gupta efficiency as about 0.017, 0.021, 0.023, and 0.91, respectively.

Abstract: While non-operative treatment with structured rehabilitation tends to be the strategy of choice in the management of Rockwood type III acromioclavicular joint injury, some advocate surgical treatment to prevent persistent pain, disability, and prominence of the distal clavicle. There is no clear consensus regarding when the surgical treatment should be indicated, and successful clinical outcomes have been reported for non-operative treatment in more than 80% of type III acromioclavicular joint injuries. Furthermore, there is no gold standard procedure for operative treatment of type III acromioclavicular joint injury, and more than 60 different procedures have been used for this purpose in clinical practice. Among these surgical techniques, recently introduced arthroscopic-assisted procedures involving a coracoclavicular suspension device are minimally invasive and have been shown to achieve successful coracoclavicular reconstruction in 80% of patients with failed conservative treatment. Taken together, currently available data indicate that successful treatment can be expected with initial conservative treatment in more than 96% of type III acromioclavicular injuries, whereas minimally invasive surgical treatments can be considered for unstable type IIIB injuries, especially in young and active patients. Further studies are needed to clarify the optimal treatment approach in patients with higher functional needs, especially in high-level athletes.

Abstract: As the popularity of Android smart devises increases, the battle of alleviating Android malware has been considered as a crucial activity with the advent of new attacks including progressively complicated evasion techniques, consequently entailing more cutting-edge detection techniques. Hence, in this paper, two Machine Learning (ML) algorithms, called Support Vector Machine (SVM) and K-Nearest Neighbors (KNN), are applied and evaluated to perform classification of the feature set into either benign or malicious applications (apps) through supervised learning process. This work involves in static analysis of apps, which checks for the presence and frequency of keywords in the Android apps' manifest file and derives the static feature sets from a 400-app dataset to produce better malware detection results. The classification performance of the ML algorithms is measured in terms of accuracy and true positive rate and interpreted to determine which algorithm is more applicable for the Android malware detection. The experimental results for a dataset of real malware and benign apps indicate the average accuracy rate of 79.08% and 80.50% with average true positive rate of over 67.00% and 80.00% using SVM and KNN, respectively.

Abstract: Traumatic anterior dislocation of the shoulder is one of the most common directions of instability following a traumatic event. Although the incidence of shoulder dislocation is similar between young and elderly patients, most studies have traditionally focused on young patients due to relatively high rates of recurrent dislocations in this population. However, shoulder dislocations in older patients also require careful evaluation and treatment selection because they can lead to persistent pain and disability due to rotator cuff tears and nerve injuries. This article provides an overview of the nature and pathology of acute primary anterior shoulder dislocation, widely accepted management modalities, and differences in treatment for young and elderly patients.

Abstract: RoboJam is a machine-learning system for generating music that assists users of a touchscreen music app by performing responses to their short improvisations. This system uses a recurrent artificial neural network to generate sequences of touchscreen interactions and absolute timings, rather than high-level musical notes. To accomplish this, RoboJam’s network uses a mixture density layer to predict appropriate touch interaction locations in space and time. In this paper, we describe the design and implementation of RoboJam’s network and how it has been integrated into a touchscreen music app. A preliminary evaluation analyses the system in terms of training, musical generation and user interaction.

Abstract: The purpose of this study is to determine the perception of financial auditor about the use of computer assisted audit techniques (CAATs) in their daily process of work. This research is a quantitative research uses, which use primary data by distributing questionnaires to the respondent. The respondents are financial auditor who worked in public accounting firm in Jakarta Special Region of Indonesia. This study tested the hypotheses between variables by using path analysis, while the independent variables in this study are Performance Expectancy, Effort Expectancy, Social Influence and Facilitating Condition. Intervening variable is Behavioral Intention and dependent variable is Use Behavior. The results of this research indicate that Performance Expectancy has significant impact on Behavioral Intention. Both Effort Expectancy and Social Influence do not have significant impact on Behavioral Intention. Facilitating Condition and Behavioral Intention have significant impact on Use Behavior

Abstract: The enhancement of the machine condition monitoring process is a key issue for reliability improvement. In fact, in order to produce quickly, economically, with high quality while decreasing the risk of production break due to a machine stop, it is necessary to maintain the equipment in a good operational condition. This requirement can be satisfied by implementing appropriate maintenance strategies such as Condition Based Maintenance (CBM) and using updated condition monitoring technologies for faults detection and classification. In this context, a new method for machinery condition monitoring based on Mel-Frequency Cepstral Coefficients (MFCCs) and Support Vector Machine (SVM) is proposed to automatically detect the mechanical faults by maximized the generalization ability. Hence, the purpose is to design an automatic detection system for mechanical components defects based on supervised classification by trained to maximize the margin. The proposed approach consists in a sequence of binary classifications after extracting a set of relevant features such as temporal indicators and MFCC coefficients. The diagnosis accuracy assessment is carried out by conducting various experiments on acceleration signals collected from a rotating machinery under different operating conditions.

Abstract: In classical feedforward neural networks such as multilayer perceptron, radial basis function network, or counter-propagation network, the neurons in the input layer correspond to features of the training patterns. The number of these features may be large, and their meaningfulness can be various. Therefore, the selection of appropriate input neurons should be regarded. The aim of this paper is to present a complete step-by-step algorithm for determining the significance of particular input neurons of the probabilistic neural network (PNN). It is based on the sensitivity analysis procedure applied to a trained PNN. The proposed algorithm is utilized in the task of reduction of the input layer of the considered network, which is achieved by removing appropriately indicated features from the data set. For comparison purposes, the PNN's input neuron significance is established by using the ReliefF and variable importance procedures that provide the relevance of the input features in the data set. The performance of the reduced PNN is verified against a full structure network in classification problems using real benchmark data sets from an available machine learning repository. The achieved results are also referred to the ones attained by entropy-based algorithms. The prediction ability expressed in terms of misclassifications is obtained by means of a 10-fold cross-validation procedure. Received outcomes point out interesting properties of the proposed algorithm. It is shown that the efficiency determined by all tested reduction methods is comparable.

Abstract: Background Massive, irreparable rotator cuff tears (RCTs) are a challenging clinical problem in young patients. In recent years, arthroscopic superior capsular reconstruction (ASCR) is a popular treatment in the massive, irreparable RCTs. However, studies reporting clinical results of ASCR are rare in the literature. Methods Between 2013 and 2015, six patients underwent ASCR. One patient treated with dermal allograft, while five patients with autogenous fascia lata graft. Demographic data, as well as preoperative and last follow-up clinical data including pain, range of motion (ROM), strength, American Shoulder and Elbow Surgeons system, the Constant system, the University of California at Los Angeles system, the Simple Shoulder Test, and the Shoulder Pain and Disability Index system were obtained. Acromiohumeral distances and Hamada classification were measured on standard anteroposterior x-ray. Results All patients were men, and the average age was 59.5 ± 4.18 years (range, 53-65 years).The minimum follow-up was 18 months with a mean follow-up was 27.33 ± 7.58 months (range, 18-36). All patients had postoperative improvement in pain scores and functional scores. The ROM and strength did not improve after surgery. The Hamada score progressed of radiographic stage in 2 patients. In the case of dermal allograft, there was graft failure 6 weeks after ASCR. Conclusions Our results support the ASCR as a viable treatment for surgical salvage in massive, irreparable RCTs. This treatment option may provide patients with decreased pain and increased function. And studying our case of dermal allograft failure provides opportunities to decrease graft failure in ASCR using dermal allograft.

Abstract: In machine learning research and application, multiclass classification algorithms reign supreme. Their fundamental property is the reliance on the availability of data from all known categories to induce effective classifiers. Unfortunately, data from so-called real-world domains sometimes do not satisfy this property, and researchers use methods such as sampling to make the data more conducive for classification. However, there are scenarios in which even such explicit methods to rectify distributions fail. In such cases, 1-class classification algorithms become the practical alternative. Unfortunately, domain complexity severely impacts their ability to produce effective classifiers. The work in this article addresses this issue and develops a strategy that allows for 1-class classification over complex domains. In particular, we introduce the notion of learning along the lines of underlying domain concepts; an important source of complexity in domains is the presence of subconcepts, and by learning over them explicitly rather than on the entire domain as a whole, we can produce powerful 1-class classification systems. The level of knowledge regarding these subconcepts will naturally vary by domain, and thus, we develop 3 distinct methodologies that take the amount of domain knowledge available into account. We demonstrate these over 3 real-world domains.

Abstract: Ultrasound has been used as one of the primary technologies utilized widely for clinical diagnosis due to its affordability, non-invasive characteristic, portability, and its fast performance in acquisition. Recently, it started to be used as a visual feedback method for tongue articulation, thanks to its capacity of real-time visualization and video capture of underlying structures inside the mouth. When an Ultrasound transducer is placed along the mid-line under a chin, it shows the tongue motion in sagittal view while speaking. As it is still quite difficult to understand the structure in ultrasound images, we proposed a guided learning system for pronunciation by visualizing tongue articulation in Ultrasound image sequences. Video image registration technique has been employed to project sagittal section of tongue back to the corresponding position on the subject head. The proposed system targets speech therapy and foreign language pronunciation lessons. Two main technology components are (i) Ultrasound tongue image segmentation and tracking (ii) registration of Ultrasound image sequences on video of a subject during the speech. Our experiments on Chinese English learners revealed that the proposed system is capable of providing the beneficial improvement on English pronunciation.

Abstract: A productive and healthy environment is directly influenced by indoor air quality parameters. Therefore, is fundamental to monitor indoor air quality environments as in a great diversity of living environments the air quality can be extremely poor. Humans typically spend more than 90% of the time indoors thus is extremely important to detect air quality problems in real-time. The unceasing scientific developments turn achievable to develop systems alongside with data collection and data sharing leading to several enhancements in ambient assisted living systems architectures. In this paper, a smartwatch-based application for enhanced living environments based on Internet of Things is presented. This system incorporate a hardware prototype for data sensing denominated by iAQ Wi-Fi and a smartwatch application that provides data consulting and notifications. The iAQ Wi-Fi incorporate wireless communication technologies and offers modularity, scalability, easy installation and smartwatch compatibility. The real-time monitoring data is stored in a cloud service named ThingSpeak and can be accessed by a smartwatch application denominated by iAQ Watchapp which allow easier access to the living environment quality in real time. Using the iAQ Watchapp the regular can analyze the monitored data in numeric or chart form.

Abstract: As the number of tag increases in the UWB Real Time Localization System (RTLS), the positioning update frequency decreases rapidly. To solve this problem, this paper introduces a TDOA system whose coordinate update frequency is not affected by the number of tags. We also present an algorithm that simultaneously measures the coordinates of Tag and Anchor. We performed wireless time synchronization to improve measurement precision and coordinate accuracy of Tags and Anchors. The experiment result confirmed that the positioning error of about 4 meter when the time synchronization is not applied is reduced to 0.08 meter or less with wireless time synchronization.

Abstract: Acute lymphoblastic leukemia can be diagnosed through a series of tests which include the minimally invasive microscopic examination of a stained peripheral blood smear. Manual microscopy is a slow process with variable accuracy depending on the laboratorian's skill level. Thus automating microscopy is a goal in cell biology. Current methods involve hand-selecting features from cell images as inputs to a variety of standard machine learning classifiers. Underrepresented in this filed, yet successful in practice, is the convolutional neural network that learns features from fine-grained images. This paper compares the performance of a convolutional neural network model with other models to determine the validity of using whole cell images rather than hand-selected features for acute lymphoblastic leukemia classification.

Abstract: Coal mining has always been a necessary evil. We need the coal for various operations, especially electrical power generation. However mining the coal has proven to be very dangerous and has caused many accidental deaths over the years. Keeping this in mind we have designed an intelligent system which can be used on helmets of these underground coal miners and can monitor/analyze a few major hazardous parameters found in these mines in real time. This includes humidity, temperature and gas contents such as methane and sulfur dioxide. These parameters, if above a certain level, can cause choking, suffocation, flooding, gas poisoning, roof collapse or explosions. Our system can detect these parameters, analyze them in real time and alert the ground control and the worker about the situation using a buzzer. Compact sensors and radio frequency modules are used to ensure practicality.

Abstract: Background The purpose of this study was to evaluate the clinical outcomes and complications of hook plate fixation in acromioclavicular (AC) joint dislocations and distal clavicle fractures. Methods We retrospectively reviewed a series of 60 consecutive patients with hook plate fixation for AC joint dislocation (group I) and distal clavicle fracture (group II). Groups I and II had 39 and 21 patients, respectively. Clinical results were evaluated using the pain visual analogue scale (VAS), simple shoulder test, and Constant-Murley scores. In addition, subacromial erosion and stiffness were evaluated as complications. Results At the removal, the pain VAS was 2.69 ± 1.30 and 4.10 ± 2.14 in groups I and II, respectively, which were significantly different (p=0.003). The simple shoulder test score was 9.59 ± 1.60 and 7.81 ± 2.67 in groups I and II, respectively, which were also significantly different (p=0.002). Subacromial erosion was significantly more frequent in group II (14/21 patients, 66.7%) than in group I (15/39 patients, 38.5%) (p=0.037), and stiffness was also higher in group II (17/21 patients, 81.0%) than in group I (22/39 patients, 56.4%), but it was not significant. Conclusions Hook plate fixation showed good clinical and functional results for the treatment of acute unstable AC joint dislocation and distal clavicle fracture. But, in distal clavicle fractures, there are more subacromial erosion and stiffness compare with acute unstable AC joint dislocation.