Linear Regression Prediction Model Research Articles

Testing the Predictability of U.S. Housing Price Index Returns Based on an IVX-AR Model

We use ten common macroeconomic variables to test for the predictability of the quarterly growth rate of house price index (HPI) in the United States during 1975:Q1–2018:Q2. We extend the instrumental variable based Wald statistic (IVX-KMS) proposed by Kostakis, Magdalinos, and Stamatogiannis to a new instrumental variable based Wald statistic (IVX-AR) which accounts for serial correlation and heteroscedasticity in the error terms of the linear predictive regression model. Simulation results show that the proposed IVX-AR exhibits excellent size control regardless of the degree of serial correlation in the error terms and the persistency in the predictive variables, while IVX-KMS displays severe size distortions. The empirical results indicate that the percentage of residential fixed investment in GDP is fairly a robust predictor of the growth rate of HPI. However, other macroeconomic variables’ strong predictive ability detected by IVX-KMS is likely to be driven by the highly correlated error terms in the predictive regressions and thus becomes insignificant when the proposed IVX-AR method is implemented. Supplementary materials for this article, including a standardized description of the materials available for reproducing the work, are available as an online supplement.

Predicting short-term traffic flow in urban based on multivariate linear regression model

In order to overcome the problems of low accuracy and time-consuming of traditional prediction methods for short-term traffic flow in urban, a prediction methods for short-term traffic flow in urban based on multiple linear regression model is proposed. The corresponding data attributes of short-term traffic flow in urban are selected by traffic operation status, and used as the original data of traffic flow prediction. According to the selected attributes, spatial static attributes data and traffic flow dynamic attributes data are collected, and fault data are identified and repaired. A multiple linear regression model for prediction of short-term traffic flow in urban is constructed to realize the prediction of short-term traffic flow in urban. The experimental results show that, compared with other methods, the average prediction accuracy of the proposed method is as high as 98.48%, and the prediction time is always less than 0.7 s, which is shorter.

Controlling the outcome of SN2 reactions in ionic liquids: from rational data set design to predictive linear regression models.

Rate constants for a bimolecular nucleophilic substitution (SN2) process in a range of ionic liquids are correlated with calculated parameters associated with the charge localisation on the cation of the ionic liquid (including the molecular electrostatic potential). Simple linear regression models proved effective, though the interdependency of the descriptors needs to be taken into account when considering generality. A series of ionic liquids were then prepared and evaluated as solvents for the same process; this data set was rationally chosen to incorporate homologous series (to evaluate systematic variation) and functionalities not available in the original data set. These new data were used to evaluate and refine the original models, which were expanded to include simple artificial neural networks. Along with showing the importance of an appropriate data set and the perils of overfitting, the work demonstrates that such models can be used to reliably predict ionic liquid solvent effects on an organic process, within the limits of the data set.

Ultrasound Imaging of the Abdominal Wall and Trunk Muscles in Patients with Achilles Tendinopathy versus Healthy Participants.

Purpose: To compare and quantify with ultrasound imaging (USI) the inter-recti distance (IRD), rectus abdominis (RA), external oblique (EO), internal oblique (IO), transversus abdominis (TrAb), and multifidus thickness and the RA and multifidus cross-sectional area (CSA) between individuals with and without chronic mid-portion Achilles tendinopathy (AT). Methods: A cross-sectional study. A sample of 143 patients were recruited and divided into two groups: A group comprised of chronic mid-portion AT (n = 71) and B group composed of healthy subjects (n = 72). The IRD, RA, EO, IO, TrAb, and multifidus thickness, as well as RA and multifidus CSA, were measured by USI. Results: USI measurements for the EO (p = 0.001), IO (p = 0.001), TrAb (p = 0.041) and RA (p = 0.001) thickness were decreased as well as IRD (p = 0.001) and multifidus thickness (p = 0.001) and CSA (p = 0.001) were increased for the tendinopathy group with respect the healthy group. Linear regression prediction models (R2 = 0.260 − 0.494; p < 0.05) for the IRD, RA, EO, and IO thickness (R2 = 0.494), as well as multifidus CSA and thickness were determined by weight, height, BMI and AT presence. Conclusions: EO, IO, TrAb, and RA thickness was reduced and IRD, multifidus thickness and CSA were increased in patients with AT.

Estimación in vivo de la canal porcina por el método de ultrasonografía: Un Enfoque de la Ecointensificación en Bioeconomia Porcina

En este estudio se analiza la relación entre el espesor de grasa subcutánea (EGS), espesor muscular (EM) y espesor de grasa subcutánea y espesor muscular juntos (EGSM), medidos con ultrasonido en un solo punto anatómico de cerdos vivos y los parámetros más importantes de la canal . En nuestro experimento utilizamos cerdos híbridos de las razas Pietrain, Landrace y Yorkshire en cantidad de 25 animales de ambos sexos. Los resultados del estudio sugieren que los parámetros de la composición de la canal aquí analizados están altamente correlacionados con el espesor de la grasa subcutánea medida con el ultrasonido en la última costilla del musculo longuísimo dorsal en el lado izquierdo del cerdo vivo (r = 0.53 - 0.67). La correlación múltiple del modelo de predicción de regresión lineal entre los parámetros de la canal y el espesor de grasa dorsal fue (0.74) y el coeficiente de determinación mostró suficiente capacidad de predicción (R2 = 0.55). Los modelos de predicción de regresión lineal de los componentes de la canal con el espesor muscular obtuvieron baja capacidad de predicción (R2 = 0.23). Igualmente en el modelo de regresión del espesor de grasa subcutánea y espesor muscular juntos ( R2= 0.25). Entre las correlaciones más importantes encontradas están: lomo derecho, carne clase A (r = 0.80), peso vivo y canal entera (r = 0.85) paleta derecha y carne clase B (r = 0.88) siendo este patrón repetitivo en todas las correlaciones de las mediciones ultrasonografica y los componentes de la canal.

Prediction of marbling score and carcass traits in Korean Hanwoo beef cattle using machine learning methods and synthetic minority oversampling technique

Pricing of Hanwoo beef in the Korean market is primarily based on meat quality, and particularly on marbling score. The ability to accurately predict marbling score early in the life of an animal is extremely valuable for producers to meet the requirements of their target market, and for genetic selection. A total of 3989 Korean Hanwoo cattle (2108 with 50 k SNP genotypes) and 45 phenotypic features were available for this study. Four machine learning (ML) algorithms were applied to predict six carcass traits and compared against linear regression prediction models. In most scenarios, SMO was the best performing algorithm. The most and least accurately predicted traits were carcass weight and marbling score with correlation of 0.95 and 0.64 respectively. Additionally, the value of using a synthetic minority over-sampling technique (SMOTE) was evaluated and results showed a slight improvement in the prediction error of marbling score. Machine Learning approaches can be useful tools to predict important carcass traits in beef cattle.

Analysis of GNSS clock prediction performance with different interrupt intervals and application to real-time kinematic precise point positioning

Continuous and timely real-time satellite orbit and clock products are mandatory for real-time precise point positioning (RT-PPP). Real-time high-precision satellite orbit and clock products should be predicted within a short time in case of communication delay or connection breakdown in practical applications. For prediction, historical data describing the characteristics of the real-time orbit and clock can be used as the basis for performing the prediction. When historical data are scarce, it is difficult for many existing models to perform precise predictions. In this paper, a linear regression model is used to predict clock products. Seven-day GeoForschungsZentrum (GFZ) final clock products sampled at 30 s are used to analyze the characteristics of GNSS clocks. It is shown that the linear regression model can be used as the prediction model for the satellite clock products. In addition, the accuracy of the clock prediction for different satellites are analyzed using historical data with different periods (such as 2 and 10 epochs). Experimental results show that the accuracy of the clock with the linear regression prediction model using historical data with 10 epochs is 1.0 ns within 900 s. This is higher accuracy than that achieved using historical data of 2 epochs. Finally, the performance analysis for real-time kinematic precise point positioning (PPP) is provided using GFZ final clock prediction results and state space representation (SSR) clock prediction results when communication delay or connection breakdown occur. Experimental results show that the positioning accuracy without prediction is better than that with prediction in general, whether using the final clock product or the SSR clock product. For the final clock product, the positioning accuracy in the north (N), east (E), and up (U) directions is better than 10.0 cm with all visible GNSS satellites with prediction. In comparison, the 3D positioning accuracy of N, E, and U directions with visible GNSS satellites whose prediction accuracy is better than 0.1 ns using historical data of 10 epochs is improved from 15.0 cm to 7.0 cm. For the SSR clock product, the positioning accuracy of N, E, and U directions is better than 12.0 cm with visible GNSS satellites with prediction. In comparison, the 3D positioning accuracy of N, E, and U directions with visible GNSS satellites whose prediction accuracy is better than 0.1 ns using historical data of 10 epochs is improved from 12.0 cm to 9.0 cm.

P1611Non-invasive predictors of LV biopsy-obtained creatine kinase activity in patients with non-failing and failing myocardial hypertrophy

Abstract Introduction Myocardial creatine kinase (CK) activity represents an important metabolic reserve: it correlates closely with contractile reserve and post-ischaemic function, keeps cytosolic [ADP] low, and optimises the free energy for ATP hydrolysis. It may also contribute to the transition to failure in the heart hypertrophied by chronic haemodynamic overload; modest up-regulation has been shown to be cardio-protective. Total CK activity measurement requires chemical freeze-and-extract methods, which destroy tissue, precluding repeated measures. To date, non-invasive assessments of human myocardial CK flux (calculated as kf × [PCr], where kf is the pseudo-first-order forward rate constant measured by 31P-magnetic resonance spectroscopy (MRS) have not assessed creatine content or total CK activity. Thus we aimed to validate kf measurement against total CK activity and investigate predictors of CK activity. Methods 39 subjects (median age 71, range 43–84) undergoing clinically indicated cardiac surgery had CK total activity measured from LV biopsy. 31 had severe AS (10 with impaired LVEF); 2 had severe primary MR; 5 had severe mitral stenosis (2 with impaired LVEF) and 1 had an LA mass. 35 of 39 contributed triplicate datasets: CK total activity, kf (31P-MRS TRiST sequence at 3T) and LV volumes (cine-MRI, 3T Siemens). 27 had severe AS (8 with impaired LVEF); other groups were the same. Exclusion criteria were prior myocardial infarction and flow-limiting coronary disease. Flash-frozen LV biopsies obtained within 15 min of cardiopulmonary bypass were analysed for CK total activity, total creatine, and citrate synthase (CS) activity (a marker of oxidative phosphorylation capacity). Results Multiple, novel correlations were observed between CK total activity (IU/mg protein) and CS activity (r=0.87, p=9e-13), total creatine (r=0.59, p=8e-5), kf (r=0.42, p=0.013), total creatine × kf (r=0.64, p=4e-5), LVESVi (r=−0.52, p=8e-4), LVEF (r=0.48, p=0.002) and LVMi (r=−0.42, p=0.009) (Panels A-E) (LVEDVi and non-indexed counterparts were also significant correlates.) The most predictive linear regression model incorporating elements of the CK rate equation included total creatine (nmol/mg protein), kf (/s), and kf × LVESVi (ml/m2) (adjusted R2=0.56, beta=0.426, 0.618, −1.968 and p=0.001, 5e-5, 0.003 respectively, Panel F). Figure 1 Conclusions These results are the first evidence of agreement between non-invasive estimates of human cardiac CK activity (kf) and freeze-extracted chemical methods. The key original finding is that it is feasible to attempt to predict CK capacity in vivo by using a combination of techniques (creatine by 1H MRS, CK kf by 31P-MRS and LVESVi by cine imaging). The key insight is that CK capacity is best estimated not simply from what the rate equation would predict (creatine and kf), but that other factors relating to failing contractility reduce CK activity independently from creatine content and enzyme kf. Acknowledgement/Funding British Heart Foundation Clinical Research Training Fellowship (FS/15/80/31803) and Programme Grant (RG/18/12/34040).

Linear Regression Models to Estimate Exchangeable Sodium Percentage and Bulk Density of Salt Affected Soils in Sahl El-Hossinia, El-Sharkia Governorate, Egypt

ABSTRACTThe objective of this study was to develop a Linear Regression Model for the prediction of soil bulk density based on organic matter content (OM) and textural fractions (% sand, silt and clay) as well as the soil exchangeable sodium percentage (ESP) based on soil sodium adsorption ratio (SAR) in some salt affected soils of Sahl El-Hossinia, El-Sharkia Governorate, Egypt. For this purpose, 160 samples were randomly taken from top of the surface soil (0–30 cm) from different locations and samples were subjected to various analyzes. XLSTAT Version 2016.02.27444 software was used to build and test conceptual and empirical models. The statistical results of the study indicated that to predict soil bulk density (BD) based on organic matter content and textural fractions the Multiple linear regression model BD = 1.817–0.730 × OM – 0.002 × Clay – 0.001 × Silt with R2 = 0.794. On the other hand, to predict soil ESP based on SAR the linear regression model ESP = 5.577 + 0.851 × SAR with R2 = 0.773. A Linear Regression Model for prediction of BD and ESP of Sahl El-Hossinia, El-Sharkia Governorate, Egypt, can be used with high prediction.

Enhanced Optimal Portfolios - A Controlled Integration of Quantitative Predictors

The intuitiveness and practicability of mean-variance portfolios largely depends on the accuracy of moment estimates, which are subject to large estimation errors and conditional on time. We propose a model accounting for factor dynamics in a Bayesian setting, in which the impact of estimation accuracy on the posterior distribution is endogenously derived from a linear predictive regression model. Thereby, we capture upside return potential for periods of high factor explained variance whilst constraining downside risk for periods of low predictive quality. Results are robust in a simulation and empirical setting.

Nondestructive Assessment of Egg Freshness using a Synchronous Fluorescence Spectral Technique

Freshness is an important index of egg quality. In this study, a synchronous fluorescence spectral technique was employed to determine the freshness of an intact egg. Synchronous fluorescence spectra of intact eggs were acquired using a fluorescence spectrometer supported by a laboratory fluorescence acquisition device and egg freshness (Haugh Unit) was obtained using destructive methods. Eggs feature fluorescence signals were mainly concentrated in two regions: A (excitation wavelength of 290 nm over the emission wavelength range of 320-380 nm) and B (excitation wavelength range of 380-570 nm over the emission wavelength range of 610-735 nm). The two regions were selected as regions of interest, which include 2581 Excitation-Emission (Ex-Em) wavelengths; stepwise discrimination analysis was performed on the 2581 Ex-Em wavelengths to choose optimal Ex-Em wavelength combinations. A Multiple Linear Regression (MLR) prediction model was built using fluorescence signals based on the optimal Ex-Em wavelength combinations. The results revealed that the freshness of an egg could be accurately predicted with Rp2 of 0.8879 and a root mean square error estimated by validation (SEP) of 6.2896. This work demonstrates that the synchronous fluorescence spectral technique has high potential for nondestructive sensing of egg freshness.

Linking the Open Water Area of the North Open Water Polynya to Climatic Parameters using a Multiple Linear Regression Prediction Model

ABSTRACTThe North Open Water polynya is a recurring polynya and is a significant feature of northern Baffin Bay in the winter. An area of open water remains as the sea ice grows around it in the fall, and in the spring the polynya gradually expands as the thinner ice surrounding it melts. It is an important overwintering area for various bird and animal species. With climate change affecting the Arctic in particular, this polynya has seen significant transformation in terms of its area of open water. Our study analyzes this transformation and attempts to understand it by linking it to climatic parameters using a multiple regression model. Using data from ice analyses from Environment and Climate Change Canada’s Canadian Ice Service, a significant difference in the annual maximum polynya extent was observed before and after 1998. The main climate indices related to the extent of the polynya for the years prior to 1998 are surface wind speed and sea-ice concentration. For the years after 1998, the main indices are surface wind speed, sea surface temperature, and surface air temperature. This study indicates that the extent of open water in the polynya could be a good indicator of a new climate regime after 1998.

Prediction of soil organic matter using an inexpensive colour sensor in arid and semiarid areas of Iran

Soil organic matter (SOM) plays a major role in agricultural and ecological processes. For this reason, accurate quantification of SOM is important for precision agriculture and environmental management. Inexpensive sensor technology could be a potential approach to achieving the accurate prediction of SOM. The objective of this study was to evaluate inexpensive colour sensor (Nix™ Pro) data for prediction of SOM in arid and semiarid areas of Iran. A total number of 85 and 152 soil samples from the soil surface (0–20 cm) were collected from the Semnan (arid area) and Qazvin (semiarid area) regions respectively. The nonlinear random forest (RF) method and linear regression were conducted to predict SOM using NixTM pro colour sensor data. The partial least-squares approach was also utilised to reduce the dimensions of the dataset, decrease the number of input variables and avoid multi-collinearity. Soil colour was measured in moist and dry conditions. Root mean square error (RMSE), correlation coefficient (r), r-square (R2), mean square prediction error (MSPE) and ratio of performance to interquartile distance (RPIQ) were used to assess the RF and the linear regression models for prediction of SOM. Moist sample data was used for estimation of the SOM because of the larger correlation between SOM and colour sensor data in moist than dry samples. In estimation of SOM, the RF model represented lower dispersion between the estimated and the actual values of SOM (RMSE = 0.42, 0.43, RPIQ = 2.2, 2.06 and MSPE = 0.19, 0.19 in semiarid and arid regions respectively). In contrast, more dispersion was obtained by applying the linear regression model (RMSE = 0.61 and 0.51, RPIQ = 1.47 and 1.76, and MSPE = 0.39 and 0.26 in semiarid and arid regions respectively). The RPIQ values for linear regression in arid and semiarid areas were 1.76 and 1.47 respectively. Hence, the use of a linear regression model for prediction of SOM in arid areas would result in acceptable reliability; however, its utilisation should be avoided in semiarid areas due to less reliable results.

Prediction of scores in pharmacology in the second MBBS university examination based on previous academic performance and gender of the students: A pilot study

Background: Academic performance is an important reason for the stress and stress due to poor academic performance further deteriorates the academic performance of the student and it became a vicious cycle. Some predictors can be identified which can help finding the student who can be considered as potentially poor performer for future and such students can be mentored. Aims and Objectives: This study was designed to predict the scores of pharmacology in final university examination based on the first MBBS score, first internal assessment score, and gender. Materials and Methods: This record-based study was done at the Department of Pharmacology, GMERS Medical College, Dharpur (Patan), Gujarat. The study was done in two steps - in step one, the prediction equation for total marks in pharmacology in the second MBBS university examination was derived through multiple linear regression on the basis of independent variables (first MBBS total marks, first internal total marks, and gender) and in the second step, the equation so obtained was tested on the second batch of the students by comparing predicted marks with the actually obtained marks in pharmacology in the second M.B.B.S. university examination. Results: In the first batch, total marks in the 1st year MBBS examination and total marks in the first internal assessment were moderately correlated with the total marks of pharmacology in the 2nd year MBSS (r = 0.605, P = 0.00 and r = 0.589, P = 0.00, respectively) and gender was not significantly contributing to multiple linear regression model for prediction of score. A predictive equation was derived and tested on the second batch of the students. It can be observed that for around 94% of the students’ actual marks lies within plus or minus 10 marks of predicted marks. Conclusion: Prediction of academic performance from this technique can be a good technique to select the potential poor performers so that these students can be targeted for extra teaching or teaching by different methodology.

SLA-Aware and Energy-Efficient VM Consolidation in Cloud Data Centers Using Robust Linear Regression Prediction Model

Virtual machine (VM) consolidation provides a promising approach to save energy and to improve resource utilization in data centers. However, the aggressive consolidation of virtual machines may lead to service-level agreements (SLA) violation, which is essential for data centers and their users. Therefore, it is very meaningful to strike a tradeoff between power efficient and reduction of SLA violation level. In this paper, we propose a host overloading/underloading detection algorithm and a new VM placement algorithm based on our proposed robust simple linear regression prediction model for SLA-aware and energy-efficient consolidation of virtual machines in cloud data centers. Different from the native linear regression, our proposed methods amend the prediction and squint toward over-prediction by adding the error to the prediction; in this paper, we propose eight methods to calculate the error. We evaluate our proposed algorithms by extended the CloudSim simulator using PlanetLab workload and random workload. The experimental results show that our proposed model can reduce SLA violation rates by at most 99.16% and energy consumption by at most 25.43% for the real-world workload.

Tarsal Tunnel Mechanosensitivity Is Increased in Patients with Asthma: A Case-Control Study.

Background: Based on changes in lung function and musculoskeletal disorders in patients with asthma, this study aimed to compare the tarsal tunnel and fibular bone pressure pain thresholds (PPTs) of patients with asthma and healthy matched-paired controls. Methods: A case-control study was performed. One hundred participants were recruited: 50 asthma patients and 50 healthy matched-paired controls. Bilaterally, tarsal tunnel and fibula bone PPTs were registered. Results: Statistically significant differences (p < 0.01) were shown bilaterally for tarsal tunnel PPT. With the exception of fibula PPT (p > 0.05), asthma patients presented less tarsal tunnel PPT than healthy participants. Statistically significant differences (p < 0.05) were shown for two linear regression prediction models of the right (R2 = 0.279) and left (R2 = 0.249) tarsal tunnels PPTs as dependent variables, and based on sex, group, contralateral tarsal tunnel PPT and ipsilateral fibula PPT as independent variables. Conclusions: The study findings showed that a bilateral tarsal tunnel mechanosensitivity increase is exhibited in patients diagnosed with asthma. The presence of asthma may bilaterally predict the PPT of tarsal tunnel. These findings may suggest the presence of central sensitization in asthma patients, which could clinically predispose them to musculoskeletal disorders, such as tarsal tunnel syndrome.

Comparison between ANN and Multiple Linear Regression Models for Prediction of Warranty Cost

Nowadays, warranty has its own priority in business strategy for a manufacturer to protect their benefit as well as the intense competition between the manufacturers. In fact, warranty is a contract between manufacturer and buyer in which the manufacturer gives the buyer certain assurances as the condition of the property being sold. In industry, an accurate prediction of warranty costs is often counted by the manufacturer. Underestimation or overestimation of the warranty cost may have a high influence to the manufacturers. This paper presents a methodology to adapt historical maintenance warranty data with comparison between Artificial Neural Network (ANN) and multiple linear regression approach.

A unified test for predictability of asset returns regardless of properties of predicting variables

Some unified tests have been proposed recently in the literature for testing predictability of asset returns based on a simple linear predictive regression model, which has a drawback that predicted variable cannot be stationary if the predicting variable is nonstationary. To solve this issue, this paper includes the difference of the predicting variable into the simple linear predictive regression. Furthermore, a unified empirical likelihood inference is developed to test the predictability regardless of the properties of the predicting variable. A simulation study is conducted to confirm the efficiency of the proposed methods before applying to a real example.

Artificial neural networks and linear regression prediction models for social housing allocation: Fuel Poverty Potential Risk Index

Fuel poverty is a pertinent issue for vulnerable households both in industrialized and developing countries, which is related to energy prices and accessibility of energy services. This research explores the feasibility of predictive models to prevent fuel poverty through the Fuel Poverty Potential Risk Index (FPPRI). Two statistical models, multiple linear regression (MLR) and artificial neural networks (ANN), have been developed and applied to predict the probability of low-income households falling into fuel poverty when being allocated a social dwelling. The case study used to validate the model is located in the Bio-Bio Region of Chile and the households considered belong to the most vulnerable social strata. The models have considered the design and constructive features of common typologies of Chilean social dwellings, family income levels, changes in energy usage patterns and energy prices. Through extensive simulation and testing, ANNs have been found to be more accurate than MLRs for all situations, with a R2 coefficient above 99.6% and 80.7% respectively, despite their greater complexity. The result of this research can be useful in providing tools to fairly and accurately assign social dwellings to vulnerable households to prevent them from falling into fuel poverty.

Quantitative screening of serum protein biomarkers by reverse phase protein arrays.

Screening biomarkers in serum samples for different diseases has always been of great interest because it presents an early, reliable, and, most importantly, noninvasive means of diagnosis and prognosis. Reverse phase protein arrays (RPPAs) are a high-throughput platform that can measure single or limited sets of proteins from thousands of patients' samples in parallel. They have been widely used for detection of signaling molecules involved in diseases, especially cancers, and related regulation pathways in cell lysates. However, this approach has been difficult to adapt to serum samples. Previously, we developed a sensitive method called the enhanced protein array to quantitatively measure serum protein levels from large numbers of patient samples. Here, we further refine the technology on several fronts: 1. simplifying the experimental procedure; 2. optimizing multiple parameters to make the assay more robust, including the support matrix, signal reporting method, background control, and antibody validation; and 3. establishing a method for more accurate quantification. Using this technology, we quantitatively measured the expression levels of 10 proteins: alpha-fetoprotein (AFP), beta 2 microglobulin (B2M), Carcinoma Antigen 15-3(CA15-3), Carcinoembryonic antigen (CEA), golgi protein 73 (GP73), Growth differentiation factor 15 (GDF15), Human Epididymis Protein 4 (HE4), Insulin Like Growth Factor Binding Protein 2 (IGFBP2), osteopontin (OPN) and Beta-type platelet-derived growth factor receptor (PDGFRB) from serum samples of 132 hepatocellular carcinoma (HCC) patients and 78 healthy volunteers. We found that 6 protein expression levels are significantly increased in HCC patients. Statistical and bioinformatical analysis has revealed decent accuracy rates of individual proteins, ranging from 0.617 (B2M) to 0.908 (AFP) as diagnostic biomarkers to distinguish HCC from healthy controls. The combination of these 6 proteins as a specific HCC signature yielded a higher accuracy of 0.923 using linear discriminant analysis (LDA), logistic regression (LR), random forest (RF) and support vector machine (SVM) predictive model analyses. Our work reveals promise for using reverse phase protein arrays for biomarker discovery and validation in serum samples.