Comparison of PM_(10) concentrations and metal content in three different sites of the Venice Lagoon: An analysis of possible aerosol sources

The Venice Lagoon is exposed to atmospheric pollutants from industrial activities, thermoelectric power plants, petrochemical plants, incinerator, domestic heating, ship traffic, glass factories and vehicular emissions on the mainland. In 2005, construction began on the mobile dams （MOSE）, one dam for each channel connecting the lagoon to the Adriatic Sea as a barrier against high tide. These construction works could represent an additional source of pollutants. PM10 samples were taken on random days between 2007 and 2010 at three different sites： Punta Sabbioni, Chioggia and Malamocco, located near the respective dam construction worksites. Chemical analyses of V, Cr, Fe, Co, Ni, Cu, Zn, As, Mo, Cd, Sb, T1 and Pb in PM10 samples were performed by Inductively coupled plasma- quadrupole mass spectrometry （ICP-QMS） and results were used to identify the main aerosol sources. The correlation of measured data with meteorology, and source apportionment, failed to highlight a contribution specifically associated to the emissions of the MOSE construction works. The comparison of the measurements at the three sites showed a substantial homogeneity of metal concentrations in the area. Source apportionment with principal component analysis （PCA） and positive matrix factorization （PMF） showed that a four principal factors model could describe the sources of metals in PM10. Three of them were assigned to specific sources in the area and one was characterised as a source of mixed origin （anthropogenic and crustal）. A specific anthropogenic source of PM10 rich in Ni and Cr, active at the Chioggia site, was also identified.

Robust Recommendation Algorithm Based on Kernel Principal Component Analysis and Fuzzy C-means Clustering

The existing recommendation algorithms have lower robustness in facing of shilling attacks. Considering this problem, we present a robust recommendation algorithm based on kernel principal component analysis and fuzzy c-means clustering. Firstly, we use kernel principal component analysis method to reduce the dimensionality of the original rating matrix, which can extract the effective features of users and items. Then, according to the dimension-reduced rating matrix and the high correlation characteristic between attack profiles, we use fuzzy c-means clustering method to cluster user profiles, which can realize the effective separation of genuine profiles and attack profiles. Finally, we construct an indicator function based on the attack detection results to decrease the influence of attack profiles on the recommendation, and incorporate it into the matrix factorization technology to design the corresponding robust recommendation algorithm. Experiment results indicate that the proposed algorithm is superior to the existing methods in both recommendation accuracy and robustness.